GENETIC ANALYSIS OF HOST PLANT RESISTANCE 
TO THE CASSAVA MOSAIC DISEASE
By
Yvonne Rosaline Naa Norle Lokko
BSc (Hons.) Botany with Zoology, University of Ghana. 
MSc Plant Biotechnology, University of London
This thesis is submitted to the University of Ghana, Legon in partial 
fulfilment for the requirement of the award for the Ph.D degree in Crop
Science
April 2002
'£,(,0 S , C3  3  L23 
/! c , c . I
^ 3 7 0 4 2 9
Abstract
The genetic control of resistance to the cassava mosaic disease (CMD) in some 
African landraces, their relationship with the widely used resistant genetic stock, 
clone 58308, and molecular markers associated with resistance to CMD in the 
landraces were evaluated in this study.
The F-i progenies of 54 cassava crosses and their parents (clone 58308, six 
improved clones and 15 African landraces) were evaluated in two genetic 
experiments in three environments in Nigeria. Genetic analysis revealed that additive 
gene effect was more important in the genetics of resistance to CMD among the 
landraces, while both additive and non-additive gene effecsts were important for 
clone 58308.
The segregating Fi crosses exhibited varying levels of resistance and susceptibility to 
CMD suggesting polygenic inheritance. Resistant phenotypes were detected in the 
crosses involving susceptible parents which suggests that resistance to CMD in the 
clones studied is due to recessive genes and susceptible phenotypes in crosses 
involving resistant parents suggested that the resistance genes are non allelic. 
Positive transgressive segregants were also detected in some crosses. The number 
of effective factors for resistance to CMD ranged from two to seven and was 
contributed by both parents in a cross. Significant differences in the mean distribution 
of F-i progeny disease severity scores further revealed allelic differences between 
three improved clones and some landraces, and that expression of resistance is 
influenced by the nature of the female parent.
Bulk segregant analysis (BSA) showed that an SSR marker SSR30-180 was 
associated with CMD resistance in a cross between the susceptible clone 
TMSI30555 and the resistant landrace TME7. Linkage analysis revealed that SSR30-
180, which was donated by TME7, was 14.9 cM from a putative CMD resistance 
locus, CMDRL. Marker-trait association detected by regression analysis further 
showed that markers, SSR30-180, E-ACC/M-CTC-225, SSR119-A and SSR119 
accounted for 57.41%, 22.5%, 32.24% and 37.83% of the total phenotypic variation 
respectively, for resistance to CMD in the mapping population. The results further 
showed differences between TME7 and six out of the 15 resistant clones used as 
parents and checks in the genetic experiments with respect to SSR30-180, 
suggesting that different alleles are involved in resistance to CMD. Two susceptible 
clones TME31 and TME117 however, had the marker, which suggests that SSR30- 
180 is not tightly linked to resistance, thus there is a need to saturate the map with 
more markers.
These findings are useful in the selection of cassava parental clones for resistance 
breeding.
Declaration
We declare that this work was carried out by Ms. Yvonne Lokko, of the University of 
Ghana, Ghana, in collaboration with the International Institute of Tropical Agriculture 
(IITA), Ibadan, Nigeria under our supervision.
Ms. Yvonne R.N. Lokko 
Student
Department of Crop Science 
University of Ghana
Dr. S.K. Offei 
Molecular Plant Virologist 
Department of Crop Science 
University of Ghana
Dr. Eric Danquah 
Geneticist
Department of Crop Science 
University of Ghana
Dr. A.G.O. Dixon
Plant Breeder
Crop Improvement Division
International Institute of Tropical Agriculture
Ibadan, Nigeria
Acknowledgements
This work was supported by the Rockefeller Foundation grant RF 97007 #28 and 
2000 FS 137 and managed by the International Institute of Tropical Agriculture (IITA), 
Ibadan Nigeria under its Graduate Research Fellowship programme.
First of all, I thank God my heavenly father, for the opportunity I had to pursue a PhD 
degree. I thank God for the strength he gave me, for the perseverance and 
successful completion of the thesis. I am grateful to Dr. John Amuasi, Executive 
Secretary of the Ghana Atomic Energy and Dr. George Klu, Director of the 
Biotechnology and Nuclear Agriculture Research Institute, Accra, Ghana for granting 
me study leave to pursue the programme. I am extremely grateful to Drs. Robert 
Asiedu and Alfred Dixon of IITA who encouraged me to submit the project proposal. 
My sincere gratitude goes to Dr. Margaret Quin for her encouragement and 
assistance in ensuring that I did come to Nigeria to begin work on the project. Many 
thanks to Dr. Richard Jefferson who was instrumental in obtaining an extension of 
the Rockefeller award.
I wish to express my profound gratitude to my supervisors Drs. S.K. Offei and Eric 
Danquah of the University of Ghana, and Alfred Dixon of IITA. I thank them for their 
advice and assistance throughout the programme. I am very grateful to Dr. Dixon, for 
providing the experimental materials for the work and his guidance in executing the 
experiments. My sincere gratitude goes to Dr. Kwaku Agyemang for his assistance 
with the genetic analyses and also to Dr. Sagary Nokoe for his advice on statistical 
analysis.
Dr. Melaku Gedil was a Godsend at the right time. I am extremely grateful for his 
help, encouragement and partial supervision of the project. I am grateful for the help
given by Dr. Martin Fregene of the International Centre for Tropical Agriculture CIAT, 
Colombia.
I am extremely grateful to all staff of the Cassava Breeding Programme of IITA who 
were always very helpful. Special thanks go to Mr. Gbenga Akinwale, Ibrahim and 
Kinsley of the cassava breeding programme for their assistance in data collection. 
Thanks to Paul Ilona, Peter, C.C. Okonkwo and Willie for setting up the field 
experiments. I am grateful for the assistance given by Richardson Okechukwu, 
Fisayo Kolade, Esther, Nkechi and Mrs. Karen Lawal.
My sincere gratitude also goes to Dr. Chris Okafor, Manager, Individual Training, 
IITA, and all the staff of the IITA training programme for their help. Many thanks to 
Mrs Chinyere Woods, for all her friendship and help in preparing the manuscript. I am 
extremely grateful for the editorial assistance given by Mrs. Umelo.
Profund gratitude to my friends Adebola Raji and Ida Yovo-Badejoko. We shared our 
joys and our tears, our excitements and our anxieties. I am thankful for the friendship 
I shared with my colleagues, Ego Uzokwe, Biodun Cladius-Cole, Francis Ogbe, 
Joseph Onyeka, Elizabeth Okai, Charles Kwoseh, Chiedozie Egesi, Ayo Salami, and 
all the members of International Association of Research Scholars and Fellows, IITA.
My sincere gratitude goes to Drs Asafo-Agyei, Kormawa and Larbi of IITA/ILRI in 
Nigeria, to and my friends and colleagues of the Biotechnology and Nuclear 
Agriculture Research Institute, Accra, Ghana, for their encouragement and well 
wishes.
Finally, I remain indebted to my father for all the sacrifices he made for me. I am 
thankful for brothers, my sister and my sisters-in-law. Their love and prayers were 
invaluable.
v
This thesis is dedicated to my father Dr. R.B. Lokko, my brothers and sister, and the 
loving memory of my mother, Mrs. Mercy Lokko.
Dedication
v i
Table of Contents
Abstract....................................................................................................... i
Declaration...................................................................................................iii
Acknowledgements....................................................................................... iv
Dedication.................................................................................................vi
Table of Contents.........................................................................................vii
List of Tables..............................................................................................xii
List of Plates...............................................................................................xv
List of Figures............................................................................................ xvi
Chapter One Introduction................................................................................1
Chapter Two Literature Review........................................................................6
2.1 Importance of Cassava and production constraints........................................6
2. 2 Viruses affecting cassava......................................................................... 7
2.3 Cassava mosaic disease........................................................................... 8
2.3.1 Characteristics of cassava mosaic viruses.................................................8
2.3.2. Effects of CMD on cassava.................................................................... 9
2.3.3 Control.............................................................................................. 10
2.4 Host plant resistance to the cassava mosaic disease (CMD)........................ 13
2.4.1 Mechanism of resistance.......................................................................14
2.5. Genetics of cassava............................................................................ 16
2.5.1 Genome analysis................................................................................ 17
2.5.2 Cytogenetics...................................................................................... 17
2.5.3 Hybridisation and genetic diversity..........................................................18
2.6 Quantitative genetics in cassava breeding................................................19
2.6.1 Mating designs....................................................................................19
2.6.2 Heterosis........................................................................................... 21
2.6.3. Number of genes and complementarity among genes affecting quantitative 
traits......................................................................................................... 22
2.6.4. Heritability....................................................................................... 24
2.6.5 Genetic correlation studies.................................................................... 25
2.7 Biometrical analysis............................................................................... 26
2.7.1 Analysis of mixed models in genetic studies............................................. 26
2.7.2. The diallel analysis............................................................................28
2.7.3 The number of effective factors............................................................30
2.8 Molecular marker technology and its application in cassava breeding and 
genetics.....................................................................................................32
2.8.1 Isozyme markers................................................................................. 33
2.8.2 DNA markers......................................................................................34
2.8.3 Restriction Fragment Length Polymorphism (RFLP)..................................35
2.8.4 Random Amplified Polymorphic DNA (RAPD)...........................................37
2.8.5 Variable number of tandem repeats (VNTR)............................................ 37
2.8.6 Microsatellite or simple sequence repeats (SSRs).....................................38
2.8.7 Amplified fragment length polymorphism (AFLP)....................................... 39
2.9 Determination of molecular markers associated with traits of interest.............. 40
2.9.1 Bulk segregant analysis....................................................................... 41
2.9.2 Linkage mapping.................................................................................42
2.9.3 Quantitative trait loci (QTL)...................................................................43
2.9.4 Screening for resistance gene analogs................................................... 45
2.9.5 Sequence tagged sites (STS)..............  45
2.9.6 Type of population............................................................................... 46
Chater Three Materials and Methods..............................................................48
3.1 Analysis of variance and combining ability analyses.................................... 48
3.1.1 Genetic Material..................................................................................48
3.1.2 Experimental design............................................................................ 48
3.1.3 Assessment of CMD ............................................................................51
3.1.4 Data analyses.................................................................................... 51
3.1.4.1 3X18 North Carolina Design II (NCD II) analysis...................................54
3.1.4.2 4X4 Diallel analysis........................................................................... 58
3.1.4.3 Relative importance of general and specific combing ability......................60
3.1.4.4 Heritability....................................................................................... 62
3.1.4.5 Heterosis......................................................................................... 62
3.2 Genetic relationships among the various sources of resistance to CMD .......... 63
3.2.1 Area under the disease progress curve (AUDPC) of the sources of resistance 
to CMD ..................................................................................................... 63
3.2.2 Principal component analysis (PCA) of sources of resistance to CMD.......... 64
3.2.3 Gene complementarity among various sources of resistance to CMD based on 
the distribution of the F-i disease severity scores............................................... 65
3.2.3.1 Transgressive segregation for resistance to CMD...................................65
3.2.3.2 Cochran-Mantel-Haesnzel.................................................................. 65
3.2.3.3 Number of effective factors (NE) contributing to resistance to CMD........... 65
3.3. Determination of DNA markers associated with resistance to CMD ............... 68
3.3.1 Genetic material and experimental procedure...........................................68
3.3.2 Resistance screening and selection of plants for bulk segregant analysis 69
3.3.3 DNA extraction....................................................................................69
3.3.4 DNA quantification.............................................................................. 70
3.3.4.1. Agarose gel electrophoresis.............................................................70
3.3.4.2 Estimation of DNA concentration and purity.........................................71
3.3.5 Determination of CMD virus strain causing symptoms in the mapping 
population..................................................................................................72
3.3.6 Bulk segregant analysis....................................................................... 75
3.3.6.1 Random amplified fragment length polymorphism (RAPD) analysis...........75
3.3.6.2 Amplified fragment length polymorphism (AFLP) analysis........................ 75
Ligation of adapters.................................................................................... 78
Selective amplification reactions....................................................................80
3.3.6.3 Simple Sequence Repeats (SSR) analysis............................................ 81
3.3.6.4 Sequencing gel electrophoresis.......................................................... 82
3.3.6.5 Silver staining...................................................................................83
3.3.6.6 Gel analysis, marker scoring and nomenclature................................... 84
3.3.7 DNA markers associated with CMD resistance in F-i population................... 85
3.3.7.1 Genetic linkage mapping.................................................................... 85
3.3.7.3 Quantitative trait loci (QTL) analysis..................................................... 89
3.3.8 DNA markers associated with CMD resistance in various sources of resistance
................................................................................................................ 89
Chapter Four Results......................................... 90
4.1 Analysis of variance and combining ability analyses.....................................90
4.1.1. 3X18 North Carolina Design II mating scheme......................................... 90
4.1.2. 4X4 Diallel mating scheme................................................................... 93
4.1.3 Combining ability effects of 3X18 NCD II mating scheme............................95
4.1.4 Combining ability and maternal effects in the 4X4 Diallel mating scheme.... 100
4.1.5 Heterosis for resistance to CMD in F-i crosses........................................ 103
4.2 Genetic relationships among the various sources of resistance to CMD .........104
4.2.1 Relationships among sources of resistance based on area under the disease 
progress curve (AUDPC) for CMD................................................................104
4.2.2. Relationships among sources of resistance based on principal component 
analysis (PCA) of CMD responses................................................................107
Fig 4.1: Plot of first principal component versus second principal component analysis 
of parents and checks across environments....................................................110
4.2.3 Complementarity of genes for resistance to CMD among various sources of 
resistance................................................................................................ 111
4.2.3.1 Frequency distribution of CMD severity scores in Fi progenies of various 
sources of resistance................................................................................. 111
4.2.3.2 Transgressive segregants in Fi progenies of various sources of resistance 
.............................................................................................................. 117
4.2.3.3 Cochran-Mantel-Haenzel (CMH) statistic test of association................... 119
4.2.3.4. Number of effective factors.............................................................125
4.3 Determination of DNA markers associated with resistance to CMD ...............131
4.3.1 Resistance screening and selection of plants for bulk segregant analysis.... 131
4.3.2 Genomic DNA quality and concentration................................................ 131
4.3.3 Cassava mosaic virus strains in mapping population............................... 131
x
4.5 Association of DNA markers with CMD resistance in F-i mapping population ... 142
4.5.1 Genetic Linkage analysis.................................................................... 142
Fig. 4.2 Linkage group-X of the linkage map showing the location of the resistance to 
CMD, defined as the CMDRL....................................................................... 152
4.5.2 QTL analysis.....................................................................................153
4.6 Association of DNA marker SSR30-180 with CMD resistance in other cassava 
clones..................................................................................................... 156
Chapter Five Discussion.............................................................................158
Chapter Six Conclusion...............................................................................176
References.............................................................................................. 178
Appendices.............................................................................................. 200
xi
Table 3.1 Description of cassava clones used as parents in 4X4 diallel and 3X18 
North Carolina Design II matings and checks............................................. 49
Table 3.2 Standard five point scale scoring systems for CMD (IITA, 1990)............52
Table 3.3 Nucleotide sequences of DNA primers used in polymerase chain reaction 
for the detection of cassava mosaic begomoviruses strain in Ft mapping 
population.............................................................................................74
Table 3.4. Nucleotide sequences of AFLP adapters, primers and primer extensions76
Table 3.5 Description of cassava SSR loci and their primer pairs used in genetic 
linkage analysis.....................................................................................88
Table 4.1 Analysis of variance for CMD severity at 12 weeks after planting (WAP) 
among all genotypes in 3X18 NCD II mating scheme, evaluated in three 
environments........................................................................................ 91
Table 4.2 Analysis of variance for CMD severity at 12 weeks after planting (WAP) 
among all genotypes in 4X4 diallel mating scheme, evaluated in three 
environments........................................................................................ 94
Table 4.3 Mean* CMD severity scores and GCA effects for CMD symptom severity of 
parents in the 3X18 NCD II mating scheme, evaluated in three environments...96
Table 4.4 Mean CMD severity scores and SCA effects for CMD symptom severity of 
crosses in 3X18 design II crosses in environment with significant variation for 
SCA..........................,......................................................................... 98
Table 4.5 Mean, GCA and maternal (MAT) effects for CMD symptom severity of 
parents in 4X4 diallel crosses in environments with significant variation for their 
effects................................................................................................ 101
Table 4.6 Mean CMD severity scores and SCA effects for CMD symptom severity of 
crosses in 4X4 diallel in environments with significant variation for their effects 
 102
Table 4.7 Area under the disease progress curve (AUDPC) of cassava clones based 
on mean CMD severity scores and their ranks for CMD resistance............... 105
Table 4.8 Spearman's rank correlation coefficient for AUDPC values for shoot-tip and 
whole plant CMD symptom severity.........................................................106
Table 4.9 Eigenvectors and eigenvalues as a proportion of total variance of the first 
three principal components for the CMD resistance responses of the 23 parental 
and check clones................................................................................. 108
List of Tables
Table 4.10. Frequency distribution of Ft disease severity scores (based on mean 
disease severity scores of progenies) in an environment and mean number of 
progenies (N) in crosses involving susceptible parents................................112
Table 4.11 Frequency distribution of F-i disease severity scores (based on mean 
disease severity scores of progenies) in an environment and mean number of 
progenies (N) in crosses involving resistant parents....................................113
Table 4.12 Frequency distribution of Fi disease severity scores (based on mean 
disease severity scores of progenies) in an environment and mean number of 
progenies (N) in crosses involving resistant and susceptible parents............ 115
Table 4.13 Frequency of positive transgressive in percentages (PTS) with disease 
severity scores at least one score better than the better parent in a cross in three 
environments.......................................................................................118
Table 4. 14 Chi-square values for differences in the mean distribution of disease 
severity scores of F-i progenies in crosses involving resistant male parents when 
crossed with each of the female parents 130001 resistant to CMD, I30555 
susceptible to CMD and I30572 resistant to CMD...................................... 120
Table 4.15 Chi-square values for differences in the mean distribution of disease
severity scores of F-i progenies in crosses involving resistant and susceptible 
male parents when crossed with each of the female parents 130001 resistant to 
CMD, I30555 susceptible to CMD and I30572 resistant to CMD................... 122
Table 4. 16 Chi-square values for differences in the mean distribution of disease 
severity scores of Fi progenies in crosses involving susceptible male parents 
when crossed with each of the female parents 130001 resistant to CMD, I30555
susceptible to CMD and I30572 resistant to CMD...................................... 124
Table 4.17 Number of effective factors (NE), number of effective factors contributed 
by better parent (NBP) and number of effective factors contributed by poorer 
parent (NPP) in F1 crosses between resistant parents................................ 126
Table 4.18 Number of effective factors (NE), number of effective factors contributed 
by better parent (NBP) and number of effective factors contributed by poorer 
parent (NPP) in F1 crosses between resistant and susceptible parents 128
Table 4.19 Number of effective factors (NE), number of effective factors contributed 
by better parent (NBP) and number of effective factors contributed by poorer 
parent (NPP) in F1 crosses between susceptible parents............................ 130
Table 4.20 Reaction of Ft plants in mapping population to CMD severity in their shoot 
tips (MST), whole plant (MWP), mean severity of first 10 leaves plant (MLS), total 
mean response presence or absence of ACMV virus and their classification 
(Class)............................................................................................... 132
Table 4.21 Average number of bands, level of polymorphism among parents (TMS 
I30555, TME7) and bulks with different AFLP systems................................139
Table 4.22 Segregation analysis for conformity to 1:1 and 3:1 segregation ratio of 
polymorphic DNA markers scored in mapping population contributed by P1 
I30555............................................................................................... 146
Table 4.23 Segregation analysis for conformity to 1:1 segregation ratio of 
polymorphic DNA markers scored in mapping population contributed by P2 
TME7.................................................................................................148
Table 4.24 Molecular markers mapped on fourteen linkage groups of cassava 
genome.............................................................................................. 151
Table 4.25 Simple linear correlation coefficient between markers and CMD 
responses...........................................................................................154
Table 4.26 Association between markers SSR30-180 and E-ACC/M-CTC-225 with all 
CMD resistance traits (All traits), mean shoot-tip severity (MST), mean whole 
plant severity (MWP), and mean leaf severity (MLS) in mapping population.... 155
x iv
Plate 3.1 Cassava leaves, with varying levels of cassava mosaic disease 
severity................................................................................................ 53
Plate 4.1 Amplification products of section of mapping population showing the 
presence of ACMV with two ACMV specific primers.................................. 135
Plate 4.2 Amplification profile generated by some RAPD primers on, parents TMS 
I30555 and TME7, resistant bulk and susceptible bulk................................ 136
Plate 4.3 Amplification products of SSR primer 30 in parents, bulks and bulk 
members............................................................................................ 141
Plate 4.4. A cross section of AFLP products of parents TMS I30555 and TME7 and 
progenies............................................................................................143
Plate 4.5. Section of AFLP products for primer P-ACC/T-CTA...........................144
Plate 4.6. Example of multiplexing PCR products of three SSR primers, I, SSR6, II, 
SSR70, III, SSR7................................................................................. 145
Plate 4.7 Amplification product of SSR 30 on 23 cassava clones........................157
List of Plates
xv
List of Figures
Fig 4.1: Plot of first principal component versus second principal component analysis
of parents and checks across environments..............................................110
Fig. 4.2 Linkage group-X of the linkage map showing the location of the resistance to
CMD, defined as the CMDRL................................................................. 152
xvi
Chapter One 
Introduction
The cassava mosaic virus disease (CMD) is considered the most devastating 
disease of cassava in Africa (Jennings, 1977; Geddes, 1990). It is caused by the 
cassava mosaic geminiviruses, transmitted by the whitefly (Bemisia tabaci Genn) and 
spread through propagation of infected vegetative propagules. The cassava mosaic 
disease causes severe yield losses ranging from 20 to95%, and the effect of the 
disease is most severe when plants are infected at the early crop stage (Terry and 
Hahn, 1980; Otim-Nape et al., 1994a; Thresh et al., 1994a). It is estimated that total 
crop yield losses due to CMD in Africa is about $2 billion per annum (IITA, 1997).
The most effective means of controlling CMD is by cultivating resistant genotypes 
(Thresh et al., 1994b). In tackling the early disease challenges in Africa, resistance to 
CMD was obtained from a cross between cassava and its relative Manihot glaziovii 
Muller von Argau (Nichols, 1947). Progenies from the third backcross generation 
were evaluated in different African countries and clone 58308, which had edible 
storage roots and resistance to the disease, was selected in Nigeria. Clone 58308 
has been the main source of resistance used by breeders for several decades and 
has resulted in the selection of several improved cultivated cassava genotypes of the 
Tropical Manihot Selection (TMS) series, with resistance to CMD (IITA, 1973; Hahn 
et al., 1989). Some of the genotypes which are widely cultivated in Africa include 
TMS 130001, TMS I30572, TMS 14(2)1425, TMS 160142 and TMS I90257.
Despite the progress made in resistance breeding, the disease is still a threat to 
cassava production. Recent disease epidemics in East Africa, caused by a 
recombinant strain of two of the cassava begomoviruses (Zhou et al., 1997) resulted 
in losses of 2.2 billion tons losses in storage root yield that amounted to US $440
1
million (Thresh et al, 1997). Furthermore, exotic germplasm from South America are 
generally not resistant to CMD and succumb to the disease under field conditions in 
Africa (Porto et al., 1994). Thus, much of the germplasm introduced from the crop's 
centre of origin, to broaden the genetic base in Africa has been lost (Jennings and 
Hershey, 1985). There is therefore a need to increase the levels of resistance within 
the cassava genepool (Thresh et al., 1994a). To ensure that durable resistance is 
maintained within the African cassava germplasm, additional sources of resistance 
with a wider genetic base will be required to diversify resistance to the disease. This 
would allow the accumulation of desirable gene combinations that would be difficult 
for the pathogens to circumvent in the long term.
Cassava is open pollinated and in the field hybridises with itself and related wild 
species creating greater variability for different traits. Silvestre and Arraudeaus 
(1983), noted that African farmers sometimes take cuttings from spontaneous 
seedlings for their subsequent planting (cited in Lefevre and Charrier, 1993). These 
new genotypes, referred to over time as landraces, provide a wealth of new genes for 
desirable traits and form part of the crop’s genetic resource (Gulick et al., 1983; 
Hershey, 1987). Some of these African landraces, the Tropical Manihot Esculenta 
series (TME), have been identified to be resistant to CMD (Raji, 1995) and could 
serve as new sources of resistance. However, selection of parents to be incorporated 
in a breeding programme cannot be based on their performance alone. Inheritance of 
traits depends on the type of gene action that may be involved and the extent of 
diversity among the parents. Understanding the mode of gene action and inheritance 
in the various sources of resistance is important in determining the appropriate 
breeding strategy to adopt.
To efficiently utilise new sources of resistance to CMD in breeding programmes and 
to diversify resistance when transferred into elite material, it is essential to compare 
the various resistant genotypes with each other to determine if the loci for resistance
2
are similar, and how their effects complement each other to enhance resistance. 
Establishing genetic relationships between the different sources of resistance would 
facilitate the choice of parents for developing new resistant cultivars. In addition, by 
combining different genes that relate to different sources of resistance, epistatic 
interaction may be identified such that higher levels of resistance can be developed 
to protect the crop.
Resistance to CMD derived from clone 58308 has been reported to be polygenic, 
recessive and inherited largely in an additive manner (Hahn and Howland, 1972). 
Polygenic inheritance was also reported for some of the African landraces (Lokko et 
at., 1998). Hahn et at. (1980a), deduced from this that resistance to the disease must 
be attributed to the combined action of a number of loci which are linked on a 
chromosome or a set of chromosomes. Population improvement and recurrent 
selection was therefore recommended in breeding for resistance to the disease 
(Hahn et at., 1979). Interestingly, a major dominant gene responsible for resistance in 
an African landrace TME3 (known as 2nd Agric in Nigeria) has been reported (IITA, 
2000; Akano et at., 2002). Further progress in understanding CMD resistance could 
be made by identifying the individual gene or genes affecting resistance and 
determining their position on the chromosome map. These resistance genes could 
then be transfered to susceptible clones using conventional crossing coupled with 
marker assisted selection (MAS) and by genetic transformation.
The expression of CMD in different cassava genotypes is known to be influenced by 
genotype x environment interaction (GXE) on the trait (Fargette et at., 1994). 
Significant GXE effect on a trait reduces the reliability of the phenotype as an 
indicator of the genotype (Kang, 1998). To obtain more precise genetic information 
on resistance to CMD, experiments are conducted in different environments (i.e. 
locations x years). This would demonstrate the GXE effect on the expression of the 
trait.
3
The detection of multiple genes for virus resistance using segregation analysis alone 
is not efficient because of the differences due to genotype by environment interaction 
(McMullen and Louie, 1989). Molecular markers which are not affected by 
environmental conditions and are insensitive to gene interactions allow geneticists 
and to locate and follow the numerous interacting genes that determine a complex 
trait as well as tagging those controlled by single genes (Botstein, et al., 1980). The 
two main strategies used to identify molecular markers associated with traits of 
interest are genetic linkage mapping and bulk segregant analysis (Tanksley et al., 
1989; Michelmore et al., 1991; Giovannoni et al., 1991).
Genetic linkage mapping as a tool for localising and isolating both simple and 
complex traits such as resistance to CMD can provide a more direct method for 
selecting desirable genes via their linkage to easily detectable molecular markers 
(Tanksley et al., 1989). Once the trait is identified and mapped, MAS could be used 
to verify its introduction into new genotypes developed in a breeding programme. 
MAS can reduce breeding population sizes, continuous recurrent testing, and the 
time required in developing a superior line by selecting the desired genotype early in 
the breeding programme.
Bulk segregant analysis (BSA) is a rapid method for identification of markers in 
specific regions of the genome (Michelmore et al., 1991; Giovannoni et al., 1991). 
The technique has the potential of detecting unique regions in the genome 
associated with the trait of interest. DNA samples are pooled from individuals with 
extreme phenotypes for a particular trait, or marker allele(s), and are then screened 
to determine molecular markers associated with the trait of interest. A marker of 
interest can then be converted into a sequence characterised amplified region 
(SCAR) marker, for the rapid screening of the trait in other populations.
4
A genetic linkage map of cassava has been constructed at the Centro Internacional 
de Agricultura Tropical (CIAT) in Colombia (Fregene et al., 1997). The fully saturated 
map would allow cassava breeders and geneticists to identify quantitative trait loci 
(QTLs) and clone the genes associated with resistance to CMD. Another map that is 
being developed at the International Institute of Tropical Agriculture (IITA) in Nigeria, 
to identify genes for resistance to CMD, could be used together with the CIAT map in 
developing a consensus map of cassava. A consensus genetic linkage map would be 
of tremendous importance to cassava research in that it would allow researchers to 
(i) identify informative markers to use in experiments, (ii) identify QTL of economic 
traits, (iii) generate a genomic database of cassava and (iv) could also be used for 
comparative mapping with other species. To facilitate saturating the maps with 
molecular markers associated with CMD resistance, more segregating populations 
need to be developed and evaluated in different environments to accumulate 
quantitative data on resistance to CMD. Bulk segregant analysis (BSA) and linkage 
analysis could then be used to place the markers on the maps.
The objectives of this study were therefore to:
i) evaluate the relative importance of general and specific combining ability, estimate 
heterosis and heritability, and determine maternal effects for resistance to CMD:
ii) evaluate the genetic relationships among the various sources of resistance to 
CMD;
iii) identify molecular markers linked to genes for resistance to CMD.
5
Chapter Two 
Literature Review
2.1 Importance of Cassava and production constraints
The cassava crop is a perennial shrub belonging to the Euphorbiaceae family. It was 
introduced into Africa and Asia from South America by Portuguese travellers in the 
fifteenth century (Jennings and Hershey, 1985; Allem, 1994).
Cassava is cultivated mainly for its storage roots, which are a major staple in Africa, 
and serves as a foreign exchange earner in many countries (Dorosh, 1988; Cock, 
1985b). The storage roots are consumed directly after boiling or processed into 
various food products (Allem and Hahn, 1991). The storage roots are also processed 
into starch for industrial purposes, flour for confectionery and animal feed. The peels 
also serve as fodder for livestock (Bokanga, 1998; Mandal, 1993). The leaves which 
contain appreciable amounts of vitamins, minerals and proteins are consumed as 
green vegetables in many parts of Africa particularly in Zaire, Sierra Leone, Liberia, 
Guinea and Cameroon (Lancaster and Brooks, 1983; West et a/., 1988; Bokanga, 
1994).
The crop can grow in poor soils on marginal lands where other crops cannot grow 
and requires minimal fertiliser, pesticides and rainfall (FAO, 2000; Cock and Howeler, 
1978). It can be harvested any time from 6 to 24 months after planting, and can be 
left in the ground as a food reserve for household food security in times of famine, 
drought and war (FAO, 2000; Best and Henry, 1994; Cock, 1985a).
Global cassava production in 1999 was over 160 million tons with Africa contributing 
over 50% of the global production cultivated on about 75 million hectares of farmland 
(Thresh etai, 1994b; FAO, 2000).
6
Cassava is cultivated in diverse agroecologies in Africa including the coastal belts, 
the rain forest, the moist and dry savannah, as well as the Sahel regions (Allem and 
Hahn, 1991; IITA 1998). It has been estimated that in Africa, the annual production of
6.1 t/ha under farmer’s conditions, is a shortfall from its potential of 30.5-51 t/ha 
(Hahn et al., 1989; FAO, 2000). This shortfall in production is attributed to yield 
losses due to diseases, pest and civil strife in some African countries, which disrupt 
farming activities (FAO, 1997). The major diseases that affect the crop in Africa are, 
the cassava mosaic disease (CMD), cassava bacterial blight (CBB), and the cassava 
anthracnose disease (CAD). The cassava green mite (CGM) is also an important 
pest of the crop (Theberge et al., 1985; Yaninek et al., 1989). Yield losses in storage 
root caused by any one of these diseases or pest in isolation is estimated at over 
80% (Yaninek et al., 1989; Lozano. 1992; Thresh et al., 1997; Otim-Nape et al., 
1994a).
2. 2 Viruses affecting cassava
Cassava is affected by a number of viruses belonging to several genera and families. 
The viruses of major economic importance are, the cassava mosaic geminiviruses 
(CMG) which are begomoviruses belonging to the geminiviridae\ the cassava 
common mosaic virus (CsCMV) which is a potexvirus; the cassava vein mosaic virus 
(CsVMV) belonging to the caulimoviridae family; the cassava frog skin virus (CsFSV) 
belongs to the redoviridae family (Bock and Harrison, 1985; Fauquet and Fargette, 
1990; de Kochko et al., 1998; Calvert et al., 1996). The CsCMV, CsVMV and CsFSV 
are important in South America, while the cassava mosaic geminiviruses are 
important in India and in Africa (de Kochko et al., 1998; Calvert et al., 1996; Harrison 
et al., 1997b). The cassava brown streak virus (CBSV) belonging to the family 
potyviridae, genus Ipomovirus has recently been identified as a major problem in 
East Africa (Msikita et al., 2000).
7
2.3 Cassava mosaic disease
The cassava mosaic disease was first reported in East Africa in 1894 (Jennings, 
1994). It is now prevalent in many parts of Africa, India and Sri Lanka (Thresh et al., 
1994a). The disease is caused by the cassava begomoviruses and transmitted by 
the whitefly (Bemisia tabaci Genn), which feeds on young cassava leaves. The 
disease is spread through propagation of infected vegetative propagules and infected 
plants serve as a source of inoculum for further whitefly spread (Storey and Nichols, 
1938; Theberge et al., 1985). The cassava mosaic disease has been rated as the 
most important vector-borne disease of any crop in Africa (Geddes, 1990).
2.3.1 Characteristics of cassava mosaic viruses
The cassava begomoviruses are composed of two circular single-stranded DNA 
particles belonging to the family Geminiviridae (Harrison and Robinson, 1988; Hong 
et al., 1993). There are five distinct cassava mosaic geminiviruses (CMG), which 
have been identified based on their nucleotide sequences. These include the African 
cassava mosaic virus (ACMV), the East African cassava mosaic virus (EACMV), the 
Indian cassava mosaic virus (ICMV), the South African cassava mosaic virus 
(SACMV), and the Uganda variant of the East African cassava mosaic virus 
(EACMV-Ugv) (Zhou et al., 1997; Berrie et al., 1998). ACMV and EACMV have been 
reported in West Africa, East Africa, and some parts of Southern Africa while the 
ICMV occurs in India and Sri Lanka (Harrison et al., 1997a; Ogbe et al., 1999; Offei 
et al., 1999). The Ugandan variant has been reported in only East Africa and SACMV 
has also been reported in only Southern Africa. Recently, another strain of EACMV 
was reported in Cameroon (Fondong et al., 2000).
The viruses can be detected and distinguished serologically by their reaction with 
specific monoclonal antibodies (Mabs) raised against them and by the polymerase 
chain reaction (PCR) using primers based on the differences in the nucleotide
8
sequences of their genome (Thomas et al., 1986; Zhou et ai, 1997). Nucleotide 
sequence analysis and serological reactions suggest that Ugv and SACMV are 
variants of EACMV and ACMV (Zhou et al, 1997; Berrie et al., 1998). Biological 
differences, such as ease of mechanical transmission into Nicotiana benthamiana 
and optimum temperature for replication, may also be used to distinguish between 
the viruses (Robinson et al., 1984; Hong etal., 1993).
2.3.2. Effects of CMD on cassava
The distinct symptom of the disease is mosaic (Rossel et al., 1992). The symptoms 
first appear in young leaves and shoot apices after infection. When the whiteflies 
feed on the plant, the virus enters and moves down rapidly to the base of the stem. It 
is then systemically transferred into side branches and shoot apices causing 
characteristic disease symptoms (Storey and Nichols, 1938; Jos et al., 1984; Rossel 
et al., 1992). Infected stem cuttings also sprout into infected plants due to the upward 
translocation of the virus particles.
The internal anatomy of cassava plants and the test plant Nicotiana benthamiana 
infected with cassava mosaic virus, have been shown to include changes in phloem 
parenchyma and sieve tube cells (Horvat and Verhoyen, 1981; Adejare and Coutts, 
1982). Reduced stem girth, plant height and petiole length of infected plants have 
been attributed to infection (Thankappan and Chacko, 1976). In some cases, the 
infected plants become stunted (Kaiser and Teemba, 1979). Recent studies have 
shown that most plants with such symptoms of severe stunting may be due to mixed 
infection, involving both ACMV and EACMV or ACMV and UgV (Harrison et al., 
1997b; Fondong etal., 2000).
Adverse effects of the disease on the physiology of plants have also been studied. 
These include a reduction in the plant photosynthetic leaf area that affects the plant’s 
metabolism (Vuylsteike, 1981; Nian et al., 1976). Vuylsteike (1981) described the
9
adverse effects on the leaves and metabolism as an increase in chlorophylase 
production in the leaves, and an increase in amylase in the storage roots, which 
adversely affects the accumulation of starch in the storage roots. Thus at harvest, 
storage root yield, quality and starch content are reduced. The quality and nutrient 
content of the leaves is also reduced due to the reduction in leaf sizes, mosaic 
pattern and physiological changes (Alagianagalingan and Ramakrishaman,1970; 
Nian etal., 1976).
Various authors have shown a positive correlation between symptom severity and 
yield reduction (Hahn et al., 1980a; Otim-Nape et al., 1994a). Fauquet et al., (1987a) 
observed that varietal responses to the disease under the same growth conditions 
also vary. Yield losses sustained in different cassava genotypes are reported to 
range from 20 to 95% (Thankappan and Chacko, 1976; Terry and Hahn, 1980; Seif, 
1982; Fargette et al., 1988). The effect of CMD on the yield of storage roots, 
however, depends on the variety and age of the plant (Thresh et al., 1994a). Losses 
are more severe when plants are infected during their early growth stage and plants 
infected through infected clonal material sustain greater storage root yield losses 
than those infected by whiteflies (Terry and Hahn, 1980; Fargette and Vie, 1988).
Fargette et al., (1988) estimated yield losses from CMD per annum of up to 30 million 
tons out of 51 million tons production in are incurred in Cote d’Ivoire. An epidemic of 
CMD during the 1990s in Uganda caused over 2.2 billion tons storage root losses 
that amounted to US $440 million (Thresh et al., 1997). It is estimated that total crop 
yield losses due to CMD in Africa is about $2 billion per annum (IITA, 1997).
2.3.3 Control
Various control methods have been suggested for the control of CMD. Since infected 
planting material is the primary source of the viruses, most of these methods are
10
based on phytosanitation, cultural practices and the use of resistant material (Thresh 
et al, 1998b; Fauquet and Fargette, 1990; Bock, 1994).
Phytosanitation include the use of virus-free planting material derived from 
thermotherapy, and meristem culture, roguing out diseased plants from within crop 
stands and removing diseased cassava plants around an intended planting site and 
exploiting the phenomenon of reversion (Thresh et al., 1998a).
In thermotherapy, the plants are grown under temperatures of 37°C and 39°C for 28 
to 105 days, to reduces the virus concentration and incidence of the disease in some 
genotypes (Kaiser and Teemba, 1979; Kaiser and Louie, 1982). In some genotypes, 
symptoms reappear when heat treatment is stopped, whiles others can remain 
healthy for 4 to 5 months after heat treatment (Gibson, 1994).
Virus-free cassava plants are produced through meristem-tip culture (Kaiser and 
Teemba, 1979; Adejare and Coutts, 1981; Ng et al., 1992). The use of heat therapy 
before meristem-tip culturing gives a higher number of virus-free plants and also 
allows the use of larger pieces (0.5 -  0.8 mm) of the meristematic dome to 
regenerate disease-free plants (Adejare and Coutts, 1981).
Roguing is recommended at the early stage of growth before the infected plants 
serve as inoculum for further spread (Thresh and Otim-Nape, 1994). This practice, 
however, is not effective in reducing CMD under high disease pressure and among 
genotypes that do not have substantial resistance to virus infection (Thresh et al., 
1998b; Fargette et al, 1990). Fargette et al. (1990) observed that CMD spread 
between fields was more important than within fields. Thresh et al, 1998b deduced 
from this that roguing could be effective only if practised by farmers throughout whole 
localities. In the 1950s in East Africa for instance, roguing as a control measure was 
effective with strict enforcement by the local authorities (Thresh et al., 1998b; Thresh 
eta!., 1994b).
11
Reversion whereby the infected cuttings produce disease-free plants in the 
subsequent generation has also been exploited to select and produce healthy 
cuttings for CMD epidemiological studies in Cote d’Ivoire (Fargette et al., 1985; 
Fargette et al., 1988). Fargette and Vie (1995) using this method showed that 
although the healthy plant may become infected in the cause of the growth period, 
these plants have less yield loss than plants that were raised from infected material.
The recommended cultural practices to control CMD are targeted at time of planting 
and stem types. Cuttings obtained from the lower portions of the main stem have 
been found to sprout into more infected plants than cuttings from the upper portion of 
the main stem and stem branches (Njock et al., 1994). Ogbe et al., (1996), 
comparing cuttings from 6 to 8 month-old mother-plants with those from 10 to 14 
month-old mother-plants, observed more infected plants in the former. It has, 
therefore, been recommended that cuttings from the upper portion of mature plants 
be used for planting to control the disease.
Another strategy to control CMD is to avoid planting when the whitefly populations 
are high. In the rainforest and transition forest agroecologies of Nigeria, for example, 
the major peak of whitefly population is between April and June. This is followed by 
a low population of the insect between July and August and a minor peak during 
September to December (Leuschner, 1978). Fargette et al. (1994) demonstrated that 
plants are generally more susceptible to secondary infection by whiteflies during the 
first 8-12 weeks. Terry and Hahn (1980), also showed that CMD infection at two 
months greatly reduced yield of storage roots. The results of Terry and Hahn (1980) 
also showed that losses due to late infections were lower in plants derived from clean 
planting material than plants from infected cuttings. Fauquet et al. (1988) suggested 
the use of low CMD incidence or low pressure sites to produce virus free planting 
material for distribution and as experimental material.
12
Plant resistance is the most important means to control virus diseases and the 
resultant crop losses due to them (Fraser, 1990). Resistant cassava genotypes tend 
to have lower CMD spread with mild to moderate symptoms (Jennings, 1960; Otim- 
Nape et al., 1998) and the resistant genotypes also have higher storage root yield 
when compared with the susceptible genotypes (Terry and Hahn, 1980). Resistant 
varieties were used to control the spread of CMD (Cours-Darne, 1968) in 
Madagascar in the 1920s. More recently in the 1990s improved resistant cassava 
varieties were used when there was an outbreak of UgV in Uganda (Otim-Nape et 
al., 1994b).
2.4 Host plant resistance to the cassava mosaic disease (CMD)
Breeding for resistance to CMD started in the 1920s in East Africa and Madagascar 
(Jennings, 1994). Initially several cassava varieties were screened and the resistant 
ones released to farmers to stop the spread of the disease (Cours-Darne, 1968). 
There was, however, the need for higher levels of resistance. Story and Nichols 
(1938), developed an interspecific cross between M. esculenta and M. glaziovii 
Muller von Argau, which resulted in hybrids which had non-tuberous root and showed 
mild and transient symptoms. After three backcrosses to cassava, they were able to 
restore root quality and retain resistance to CMD (Nichols, 1947). Intercrosses 
between the third backcross selections produced a number of promising clones with 
good quality roots and resistance to virus was enhanced (Jennings, 1994). Open 
pollinated seeds from these clones were distributed to cassava breeders in Africa 
including Nigeria where clone 58308 was selected (Hahn et al., 1989; Jennings, 
1994). The clone 58308 has poor root yield but has maintained resistance to CMD. It 
has high parental value for CMD and has been used as the main source of 
resistance in IITA’s breeding programme resulting in several improved cultivars of the 
Tropical Manihot Selection (TMS) series, with resistance to CMD (IITA, 1973; Hahn 
et al., 1989).
13
Resistance from this source is known to be polygenic, recessive, and is additively 
inherited with a heritability of over 60% (Hahn et al., 1980a; Hahn et al., 1989). 
Therefore, population improvement and recurrent selection are recommended in 
breeding for resistance to the disease (Hahn et al., 1989). This would accumulate 
desirable gene combinations that would be difficult for the pathogens to circumvent in 
the long term.
Mutation breeding, somaclonal variation and genetic engineering have been 
suggested as additional tools in breeding for resistance to the cassava mosaic virus 
disease (Ahiabu et al., 1995; Ahloowalia, 1995; Maluszynski et al., 1995; Fauquet 
and Beachy, 1991). These methods, however, require skills in mutant induction and 
selection, generating somaclones and transgenics, and an efficient regeneration 
system to avoid chimeras and undesirable mutants.
Although no genes conferring resistance to CMD have been cloned, transformation 
protocols for cassava have been established (Schopke et al., 1997). The coat protein 
sequences of the virus have been characterised (Hong et al., 1993) and could be 
used to generate gene constructs for transformation. The defective infective (Dl) 
DNA of the ACMV-B particle has been shown to induce amelioration of symptoms in 
transgenic N. benthamiana plants and the gene was passed on when the transgenics 
were used as a source of inoculum (Stanley et al., 1990). This could also be adapted 
in cassava transformation, for resistance to CMD.
2.4.1 Mechanism of resistance
The mechanism of resistance to CMD is not fully understood and no immunity to 
CMD has been found within the Manihot species. Resistance to CMD is either by 
resistance to virus infection or resistance to the vector, B. tabaci (Rossel et al.,
1992). Cassava genotypes which are resistant to virus infection show mild 
symptoms, which may be restricted to some shoots and have higher storage root
14
yield when compared to the susceptible genotypes (Jennings, 1976; Otim-Nape et 
al., 1998; Fargette et al, 1996; Terry and Hahn, 1980). They recover from symptoms 
with age (Njock et al., 1996) and contain low virus concentrations (Fargette et al., 
1996). The ability of resistant genotypes to localise or restrict the movement of the 
virus has been referred to as non-systemicity (Thresh et al., 1998b; Njock, 1994). 
Infected cuttings of resistant genotypes are also able to sprout into healthy plants in 
subsequent generations. This phenomenon is known as reversion (Fauquet et al., 
1987b).
It has been observed that the extent of reversion is higher when plants are infected 
by the whitefly earlier in its growth cycle rather when infection is due to existing virus 
concentration in the mother cuttings used in planting or when infection by whitefly is 
delayed later in the plant’s growth cycle or (Thresh et al., 1998b). Simulation models 
have postulated that virus infection of resistant genotypes couid be maintained at an 
equilibrium below 50% by the combined effect of reversion and the use of virus-free 
planting material (Fargette et al., 1994; Fargette and Vie, 1995).
Resistance to the whitefly vector has also been established. Fauquet et al. (1987b) 
observed that there was a clearly pronounced resistance to the vector. Fargette et al. 
(1996) also reported that resistant and susceptible cassava genotypes display a wide 
range of responses to whitefly population. Legg (1994), also observed variation in the 
suitability of some cassava genotypes, as hosts of B. tabaci. In contrast, Hahn et al. 
(1980b) observed similar numbers of whiteflies on resistant and susceptible 
genotypes and inferred that resistance to the vector was unlikely.
Resistance to CMD is assessed by a number of components of resistance. These 
include field incidence (percentage of infected plants); vector resistance (number of 
adult whiteflies per plant); inoculation resistance, virus resistance (virus content 
estimated by ELISA), virus diffusion resistance (development of symptoms with time),
15
symptom intensity and non-systemicity incidence of young growing tips or shoot tips 
(Fargette et al., 1996; Changa, 1992). Fargette et al. (1996) demonstrated a 
significant correlation between symptom severity and ACMV titre among resistant 
genotypes in two separate trials. Resistance to virus multiplication and movement 
were the major responses to ACMV detected among resistant cassava landraces 
and improved cultivars (Ogbe, 2001).
2.5. Genetics of cassava
Cassava is a highly heterozygous species (Kawano, 1978; Lefevre and Charrier, 
1993). When selfed, it exhibits inbreeding depression but with outcrossing it results in 
wide segregation of most traits (Kawano, 1978). The F, is, therefore, genetically 
equivalent to an F2 with respect to loci common to parents (Magoon and Krishnan, 
1977). It is propagated using stem cuttings to reproduce the genotype since 
propagation via seeds results in trait losses in the genotype due to the highly 
heterozygous nature (Kawano, 1980). Apomixis whereby seeds are produced without 
undergoing the sexual process has also been reported (Grattapagalia et al., 1996). 
Cultivated cassava genotypes generally have erratic flowering (Bonierbale et al., 
1997a). Development of crosses for genetic studies has to some extent been difficult 
because of some levels of sterility, low seed set per pollination, and seed sterility that 
exist in cassava (Kawano, 1980; Hahn et al., 1979). Nevertheless, some genetic 
studies and intraspecific and interspecific crosses have been made for genetic 
studies and genotype improvement. Some of the well documented genetic studies in 
cassava include studies on cytogenetics of cassava and its wild relatives (Bai, 1987; 
Fregene, 1996), genome analysis and diversity studies (Awoleye et al., 1994; 
Grattapagalia et al., 1996; Lefevre and Charrier, 1993), hybridisation (Haysom et al., 
1994; Asiedu et al., 1992; Magoon 1967; IITA, 1988; Bai et al., 1993; Wanyera, 
1993) and the mode of inheritance of some morphological traits (Hahn et al., 1989; 
Bryne, 1987; Rajendran, 1989; Amma et al., 1995).
16
2.5.1 Genome analysis
Cassava generally has a diploid genome (2n=36). However, some authors have 
described it as being an ailotetraploid. Since the breeding behaviour of allotetraploids 
is similar to that of diploids (Wricke and Weber, 1986), this is not certain. Although 
most cassava genotypes studied are diploid, spontaneous polyploids such as 
triploids (3x) and tetraploids (4 x) of some genotypes have been reported (IITA, 1980; 
Hahn et al., 1989). The nucleic acid content of diploid cassava is 1.67 picogram per 
nucleus, that is, 772 mega base pairs in the haploid genome (Awoleye et al., 1994). 
Lefevre and Charrier (1993) confirmed the heterozygous nature of the cassava 
genome and its relatives in their study on isozymes diversity among Manihot 
germplasm. Their studies also confirmed the diploid genome based on the 
inheritance of the isozyme markers. Similar studies on the inheritance of molecular 
markers also indicated a diploid genome (Fregene et al., 1994).
2.5.2 Cytogenetics
Fregene (1996), reviewed the work of several authors who postulated that cassava is 
a segmental ailotetraploid with basic chromosome number x =9. Jos and Nair (1979), 
however, observed regular 18 bivalents during meiosis in several cassava genotypes 
the chromosomes suggesting a diploid genome with 2n=2x=36 chromosomes. They 
also detected the presence of pachytene abnormalities such as deletions, 
duplications, inversion and inregular pairing, in cassava clones that were partially 
pollen-sterile and concluded that this was due to hybridisation (Jos and Nair, 1979). 
Normal pairing at meiosis were observed in different independent studies of 
interspecific Fi crosses, between cassava and M. glaziovil and between cassava and 
arborescent cassava, which confirmed that interspecfic hybridisation occurs within 
the genus (Magoon et al., 1969; IITA, 1988; Bai, 1982; Wanyera, 1993).
17
2.5.3 Hybridisation and genetic diversity
There is no genetic barrier to hybridisation within the genus Manihot (Haysom et al., 
1994; Asiedu et al., 1992; Magoon, 1967; Bai, 1982; Wanyera, 1993). Spontaneous 
hybrids between cassava and other Manihot species have been reported to occur 
naturally in Africa and in Brazil (Nassar, 1994).
In earlier hybridisation studies in the late 1930s, Doughty suggested that tree 
cassava is a natural hybrid between cassava and M. glaziovii (Fregene, 1996). 
Doughty’s suggestion was later confirmed by other workers who observed normal 
pairing at meiosis in an F-i between cassava and M. glaziovii and in an F-, between 
cassava and arborescent cassava (Magoon, 1967; IITA, 1988; Bai, 1982; Wanyera, 
1993). Resistance to the cassava mosaic and the cassava bacterial blight diseases 
were derived from an interspecific cross between cassava and M. glaziovii (Nichols, 
1947; Jennings, 1977; Hahn et al., 1989).
The extent of genetic diversity in the crop may be attributed to the occasional use of 
seeds in propagation since cassava would combine easily with itself and wild 
relatives. Silvestre and Arraudeaus (1983), noted that African farmers sometimes 
take the cuttings from spontaneous seedlings for their subsequent planting (cited in 
Lefevre and Charrier, 1993). These new genotypes, referred to as the landraces 
provide a wealth of allelic genes for some traits and form part of the crops genetic 
resource which could be included in breeding programmes for the improvement of 
the crop (Gulick et al., 1983; Hershey, 1987). Some of these African landraces, 
collected in West Africa have been identified to be resistant to CMD (Raji, 1995; 
Mignouna and Dixon, 1997).
The incorporation of the African landraces into the cassava breeding programme at 
IITA was hindered for several years due to the fact that many of them exhibit shy 
flowering or no flowering at all. It was discovered in 1990 that these landraces
18
flowered at Ubiaja, in Edo State, Nigeria, and this has encouraged the development 
of several crosses for genetic improvement and for genetic studies (IITA, 1990).
2.6 Quantitative genetics in cassava breeding
2.6.1 Mating designs
Most traits studied in cassava are polygenic (Byrne, 1987; Rajendran, 1989; Amma 
et al., 1995; Hahn et al., 1989). Variation in polygenic traits is attributed to 
quantitative trait loci, QTLs. Quantitave traits in plants and animals are studied using 
a variety of genetic models and designs. The analysis of mating designs allows 
genetic analysis to be carried out after one generation and provides tests of the 
adequacy of the model (Mather and Jinks, 1982).
The diallel and the North Carolina design II (NCD II) mating designs provide genetic 
interpretations including combining abilities and the inheritance of quantitative traits 
(Kang, 1994). The concept of general and specific combining ability was first defined 
by Sprague and Tatum, (1942). Knowledge on the relative importance of general 
combining ability (GCA) and specific combining ability (SCA) which represent two 
major modes of gene action for quantitative traits is essential in formulating an 
efficient breeding strategy. GCA of a line refers to the average value of the line based 
on its performance when crossed with other lines. It is due largely to additive gene 
effect. SCA is the deviation of a cross from the average GCA of the parent lines and 
is due to non-additive gene effects (Sprague and Tatum, 1942; Falconer and 
Mackay, 1997).
The theory of the diallel was first developed by Jinks and Hayman (1953). Hayman 
(1954a; b) gave some assumptions which should be satisfied before a meaningful 
interpretation of the analysis is done. These assumptions are that:
i, the parents are homozygous,
19
ii, there is normal diploid segregation,
iii, there are no maternal effects,
iv, genes are independently distributed among the parents, and
v, there is no allelic interaction or no epistasis.
These assumptions, have been shown to be unrealistic (Kempthorne, 1956; 
Matzinger and Kempthorne, 1956; Gilbert, 1973; Feyt, 1976). Kempthorne (1956) 
considered that the assumption concerning the independent distribution of genes in 
the parents is not reliable because it implies that the presence or absence of an allele 
at a particular locus is statistically independent of the presence or absence of an 
allele at any other locus. Failure of the assumption is a result of linkage of genes. 
Gilbert (1973) suggested that because of the biochemical pathways involved in the 
expression of traits, the assumption of no epistasis cannot be justified. Feyt (1976) 
showed that genes at n loci cannot be independent unless a minmum of 2n parents 
are used in a diallel cross. In other words, if four gene loci condition a trait, these loci 
cannot be independent unless 24=16 parents are used in a cross (Kang, 1994). This 
further shows that this assumption is unrealistic. Matzinger and Kempthorne (1956) 
considered varying degrees of inbreeding as acceptable to complete homozygosity of 
the parents.
The North Carolina design II (NCD II) mating scheme is a cross-classification design 
that was first proposed by Comstock and Robinson (1948). It differs from the diallel in 
that different sets of parents are used as males and females. Thus, it accommodates 
more parents in determining combining abilities than a diallel and provides the same 
sort of genetic information (Hallauer and Miranda, 1988). The main effects of males 
and females are equivalent to GCA and the female x male interaction is equivalent to 
SCA.
20
Both the diallel and NCD II mating designs have been used to obtain genetic 
information on morphological and agronomical traits of importance in cassava (Hahn 
et al., 1989; Rajendran, 1989; Amma et al., 1995). Rajendran (1989) reported 
additive gene action for storage root yield and non-additive gene action for the yield 
components (harvest index, storage root number and storage root weight). Amma et 
al., (1995) reported that root quality traits, starch content, dry matter and cyanide 
(HCN) content are predominantly non-additive. Low HCN is also recessive (Kawano, 
1978; Amma et al. 1995).
Resistance to CMD derived from clone 58308 has been reported to be polygenic, 
recessive and inherited largely in an additive manner (Hahn et al., 1980a, 1989; IITA 
1973; Msabaha, 1983). Recently, the presence of a major dominant resistant gene in 
the African landrace TME3 (known locally in Nigeria as “2nd agric”) designated as 
CMD2was reported (Akano etal., 2002; IITA, 2000).
CBB, which was also derived from the 58308 source of resistance, is inherited largely 
in an additive manner. However, some degree of non-additive effect is also involved 
(Hahn etal., 1980a; 1989; Umemura and Kawano, 1983; Jennings, 1977).
2.6.2 Heterosis
Heterosis is defined as the amount by which the mean of an Fi exceeds its better 
parent or midparent (Kang, 1994). Heterosis occurs when directional dominance 
exists, or there is allelic interaction or overdominance at some or all alleles and when 
the parents differ in gene frequency. However, the genetic basis of heterosis is not 
quite clear (Kang, 1994; Falconer and Mackay, 1997).
Evidence from maize and other crops suggest that the genetic basis of heterosis is 
due to partial or complete dominance (Hallauer and Miranda, 1988; Brummer, 1999).
21
Overdominance arising from dominant alleles in repulsion at linked loci and epistasis, 
may also play a role in heterosis expression (Xiao et al., 1995; Yu et al., 1998).
The condition that differences in allelic frequencies should exist for heterosis to be 
expressed, suggests that crosses between diverse populations would result in 
heterosis. However, highly divergent populations or certain cases of multiple allele 
populations may prevent heterosis expression (Moll et al., 1962; Bonierbale et al.,
1993). The amount of heterosis exhibited in crosses between local and exotic 
cultivars of a crop is an important criterion in selecting cultivars to be used in a cross 
(Kang, 1994). To maximise the potential of heterosis of a trait in a breeding 
programme, crosses among the various populations, lines or clones need to be 
evaluated. If the progenies express heterosis the two parents are accorded to 
different heterotic groups (Hallauer and Miranda, 1988; Brummer, 1999).
The extent of heterosis for most traits in cassava has not been studied extensively. 
Arnma et al. (1995) detected positive and significant high parent heterosis for dry 
matter content and starch in a diallel cross involving inbred lines. They also detected 
negative and significant heterosis for HCN. Heterosis for resistance to CMD was 
observed in crosses involving some improved clones and resistant African landraces 
(Lokko et al., 1998).
2.6.3. Number of genes and complementarity among genes affecting 
quantitative traits
In Mendelian genetics, analysis of the number of genes is determined by observing 
the segregation ratios. In quantitative genetics, however, effective factors are 
estimated to determine if a few or many genes condition the quantitative trait. This is 
because quantitative genetics is based on the infinitesimal model where a large 
number of loci (polygenes) all with small effects affect the trait (Lynch and Walsh,
22
1998). Environmental effects usually modify these multiple genetic factors. The most 
important assumption for estimating effective factors is additivity of gene action.
Molecular markers or biometrical methods can be used to estimate effective factors 
for quantitative traits (Lynch and Walsh, 1998; Zeng et al., 1990). Most of the 
biometrical estimating techniques do not estimate the exact number of genes but 
rather the number of effective factors (Wright, 1968; Lande, 1981; Cockerham, 1986; 
Zeng et al., 1990; Zeng 1992). This is because the number of genes affecting 
quantitative traits do not have the same effect and the total response of a trait 
depends primarily on the number of loci, which affect it (Falconer and Mackay 1997).
The biometrical methods for estimating effective factors are based on the method 
which was first developed by Castle (1921) and Wright (1968). Using the Castle- 
Wright formula, the number of effective factors in theory cannot exceed the number 
of independent segregating chromosome segments, which is the haploid number. 
Since there are usually 1 or 2 recombination events in eukaryotes, the maximum 
number of effective factors can be 2 to 3 times the haploid chromosome number. 
However, each segment can contain many genes (Lynch and Walsh, 1998).
The number of effective factors affecting resistance to CMD has not previously been 
reported. Hahn et al. (1980a) inferred from the polygenic nature of resistance to CMD 
that resistance must be attributed to the combined action of the number of loci which 
are linked on a chromosome or a set of chromosome of a genome or genomes. 
They, however, suggested that since cassava is genetically heterozygous and 
probably an allotetraploid, a study on the genetic mechanism of resistance to CMD is 
complicated and difficult.
The main purpose of testing gene complementarity is to examine allelic relationships 
between different sources of a trait in order to determine whether the genes are 
allelic among themselves. Progenies between parents varying for a quantitative trait
23
would exhibit phenotypes ranging between those of the parents. The nature of the 
progeny’s distribution is determined by the number of genes which account for the 
differences between the parents and environment (Lynch and Walsh, 1998). Gene 
complementarity among sources of resistance to disease and pests has been 
reported in oats, soybean and sorghum (Fox et al., 1997; Wang et al., 1998; Dixon et 
al., 1991).
2.6.4. Heritability
Heritability is defined as the percentage of total phenotypic variability for a trait that is 
due to the genes, and their interaction with other genes (Kang, 1994). If the 
environmental effect on the trait were negligible, then heritability would be 100%. 
Heritability is an indication of ease with which a trait can be transferred to the 
progeny. The value is not a constant and is defined for a particular trait in a 
population at a particular time. Variation in the estimate may, therefore, occur from 
population to population or within a population from time to time due mainly to 
changes in the additive genetic variance. Two forms of heritability are useful in 
breeding; heritability in the broad sense which is the ratio of total genetic variation to 
phenotypic variation, and in the narrow sense which is the fraction of the phenotypic 
variation that is due to additive genetic variation (Falconer and Mackay, 1997).
From a seven-parent diallel cross, heritability of 60% was estimated for resistance to 
CMD (Hahn and Howland 1972; IITA, 1973). In related studies, selected clones from 
10,000 F-i crosses made by crossing 58308 with three other cassava clones, gave 
narrow sense heritability of 35% and 30% for resistance to CMD estimated by parent 
offspring regression for two separate readings (Hahn et al., 1980a). Broad sense 
heritability of 48% and 68% for resistance to CMD estimated from variance 
components of half sib families from unreplicated trials in two planting seasons of 
952 clones were also reported. Similar high heritability estimates were obtained for
24
CBB from these data sets (Hahn et al., 1980b). Based on these results, resistance to 
CMD and CBB is reported to be highly heritable (Hahn et al., 1989).
Amma et al. (1995) reported high heritab.ii.ity (76.81%) for low HCN, based on narrow 
sense heritability, which was similar to an earlier report on the heritability of low HCN 
by Kawano (1978). Dixon et al.,(1994), on the other hand, reported that heritability for 
low HCN can vary from 0 to 50% depending on the composition of genotypes 
evaluated, number of years, locations, and seasons. Iglesias and Hershey (1991) 
observed that starch and dry matter content were highly heritable traits based on 
broad sense heritability, Amma et al., (1995), however, reported low heritability 
(3.25% for dry matter and 7.58% for starch content) based on narrow sense 
heritability.
2.6.5 Genetic correlation studies
The phenotypic values of different traits are often found to be correlated. This may be 
due to the fact that genes, which control the expression of a quantitative trait, are 
likely to influence the expression of others. Genetic correlation or covariance 
between traits can be mainly attributed to pleiotropy, which is the influence of a single 
gene or locus on more than one trait, or gene linkage, which is a cause of transient 
correlation (Falconer and Mackay, 1997).
Resistance to CMD and CBB is reported to be genetically correlated (IITA, 1973; 
Hahn 1978; Hahn et al., 1980b). Hahn et al. (1980b) obtained phenotypic (r2=0.423) 
and genotypic (1^ =0.899) correlations between resistance to CMD and CBB in half- 
sib families and genetic correlation (r2=0.689) in 925 clones. They attributed the 
correlation to genetic linkage, quoting the observation by Magoon et al. (1969) of 
random transmission of parental chromosomes in Manihot interspecific hybrids in 
support. Using canonical analyses, Jennings (1977) observed that the joint 
resistance was inconsistent in a population of both M. glaziovii and cassava
25
genotypes from the 58308 stock. Jennings proposed the alternative hypothesis of 
pleiotropic effect (Jennings, 1977; Jennings and Hershey, 1985).
This reported linkage between resistance to CBB and CMD could have positive 
implications in breeding for resistance to the two diseases of cassava in that 
selection for one would lead to progress in the other.
2.7 Biometrical analysis
2.7.1 Analysis of mixed models in genetic studies
For most genetic experiments, genetic material is evaluated over a set of years and 
locations or environments, to obtain precise genetic information. In such analysis, the 
genetic materials are considered fixed and the years and locations or environmental 
effects considered random (McIntosh, 1983; Zhang and Kang 1997). For random 
effects models, inferences are based on variances while in a fixed effects model, the 
inferences are based on means (Littell et al., 1996; Kang, 1994). Mixed model 
analyses, involving both fixed and random effects, allow the proper analysis of 
experimental data involving both random and fixed effects (Littell et al., 1996). The 
analysis of variance (ANOVA), the general linear models (GLM) and the variance 
components analysis (VARCOMP) which are statistical procedures available with the 
statistical analysis system (SAS), handle random, fixed and mixed model analyses 
with specifications.
The GLM and ANOVA procedures in SAS use the method of least squares to fit 
general linear models. ANOVA handles balanced data only, while GLM handles 
unbalanced data and also allows the estimation of least square means. Least square 
means are the expected values of class means that would be expected for a 
balanced design of the class variable with all the covariates at their true mean (SAS, 
1999). Least square means are thus suitable for estimating genetic effects such as 
combining abilities and heterosis in a mating experiment. It has been argued that
26
most analyses involving fixed effects models may indeed be mixed models and that 
the conventional least squares estimates in the analysis of variance may not lead to 
the 'best' estimates and may not be as informative as it should be (SAS, 1999). 
Henderson mixed model procedure of best linear unbiased predictors (BLUPs) are 
estimated in PROC MIXED procedure in SAS (SAS, 1997; Littell etal., 1996).
The VARCOMP procedure in SAS computes estimates of the variance components 
in a general linear model. It is designed to handle models that have random effects. 
The minimum variance/norm quadratic unbiased estimate (MIVQUEO) method 
produces estimates which are locally the best quadratic unbiased estimates, given 
that the true ratio of each component to the residual error component is zero (SAS, 
1997; SAS, 1999). The technique is similar to the type I methods which equates 
mean squares involving the random effects to its expected mean square, but the 
random effects in MIVQUEO are adjusted for the fixed effects.
With fixed genetic effects in the analysis of data from either a diallel or a NCD II 
experiment, it is not appropriate to draw inferences about hypothetical populations 
(Ross et al., 1983). However, with qualifications, important quantitative information 
such as the ratio of additive variance to total genetic variance can be drawn from 
such studies that may aid in the improvement of the crop for the trait (Ross et al., 
1983; Pedersen et al., 1983; Rooney et al., 1997). The ratio of additive variance to 
total genetic variance in the population gives an indication of the relative importance 
of GCA in predicting progeny performance. The closer this ratio is to 1, the greater 
the chances of predicting progeny performance based on GCA (Baker, 1978; Kang,
1994). If the genetic model assumed is a fixed effects model, the analogous mean 
square components of the fixed effects can be estimated to compute the ratios 
(Baker, 1978; Wricke and Weber, 1986; Ross et al., 1983; Kang 1994). The 
MIVQUEO procedure is, therefore, suitable for estimating the analogous mean 
square components.
27
2.7.2. The diallel analysis
A number of genetic models and statistical procedures exist for analysing and 
interpreting the results of diallel crosses. Depending on the objectives and availability 
of materials for an experiment, reciprocal crosses could be included in a diallel 
analysis. The methods used for the analysis of a diallel cross include;
i, Hayman’s method of analysis of variance for the genetic parameters ‘a’, ‘bs’, ‘c’ 
and ‘d’ (Hayman, 1954a),
ii, Combining ability analysis which partitions combining ability into general 
combining ability (GCA) and specific combining ability (SCA) (Griffing, 1956b; 
Kempthorne and Curnow, 1961; Pooni et al., 1984; Singh and Chaudhary, 
1985).
iii, Gardener and Eberhart’s analysis II which estimates the effects vh hg, h and sy 
variety and heterosis effects (Gardener and Eberhart, 1966).
The genetic parameter 'a' of Hayman’s method refers to additive gene effect and is 
equivalent to GCA, while the ‘bs’ which refers to dominance gene effect is equivalent 
to SCA. The paramter 'c’ refers to maternal effects and ‘d’ refers to reciprocal effects 
of the differences due to 'c\ The SCA is further partitioned into 'b/ (directional 
dominance which compares the mean of the Fi and the mid parent value), ‘b2 
(directional dominance which quantifies gene distribution over arrays) and ‘b3 
(discrepancy in reciprocals due to dominance).
Using the Hayman analysis of variance, significance of the b effects allows further 
analysis of the data using the graphical analysis procedure, WrIVr graphs (Mather 
and Jinks, 1982). Here, Wr is the covariance between the parents and Vr is the array 
variances, an array being the crosses in which a particular parent is involved (Singh 
and Chaudhary, 1985). In the absence of non-allelic interaction and with independent
28
distribution of genes among parents, Wr is related to Vr by a regression slope. The 
distance between the origin and the intercept of the regression line on the curve 
provides a measure of the degree of dominance. The order of the array points along 
the regression lines gives an indication of dominant and recessive genes among the 
parents (Singh and Chaudhary, 1985).
The theories underlying the analysis of diallel crosses in relation to combining ability 
were later explained by Griffing (1956a; b). Griffing (1956a) described four methods 
of the diallel crosses depending on the material used. Method (1) includes selfed 
parents, and reciprocals. Method (2) has the selfed parents and F-|S only. Method 
(3) involves the F^ and reciprocals, and method (4) involves the F-|S only. In 
Griffing’s approaches to the analysis of diallel crosses, variance components due to 
general and specific combining ability are estimated which are then translated into 
genetic components such as additive and dominance variance. The analysis may be 
based on fixed effects model I or random effects model II. The procedure for the 
analysis of variance of the two methods is the same except for the expectation of 
mean squares (Griffing 1956b; Pooni et al., 1984; Singh and Chaudhary, 1985). The 
first part of the analysis consist of testing the null hypothesis, that there are genotypic 
differences among F-, crosses, parents and the reciprocals. Significant differences 
among the genotypes are a prerequisite for the combining ability analysis (Singh and 
Chaudhary, 1985).
The average heterosis ‘h’ in the Gardener and Eberhart analysis II is equivalent to 
the parent versus cross orthogonal contrast in a traditional analysis of variance and 
the specfic heterosis ‘s,/ is equivalent to SCA (Hallauer and Miranda, 1988; Pooni et 
al., 1984). The Gardener and Eberhart analysis II also gives two estimates of the 
parent, the variety heterosis, ‘h,/ and the variety mean square 'v, which includes 
information of the performance of the varieties themselves and the varieties within 
crosses. These two estimates are not equivalent to either the parent mean square in
29
a traditional analysis of variance or GCA of Griffing’s methods (Hallauer and 
Miranda, 1988; Pooni et al., 1984). In the analysis of variance of genetic data, an 
orthogonal contrast between the parents and crosses sums of squares is used as an 
estimate of average heterosis in a mating experiment (Hallauer and Miranda, 1988).
The various methods of analyses of diallels were evaluated by Arunachalam (1976) 
who concluded that the various assumptions made for the Hayrnan, (1954a) and the 
Gardener and Eberhart (1966) methods were generally not fulfilled in practical 
situations with experimental material. He concluded that for testing parents and Fi 
materials, the combining ability analysis gives the best estimates of the gene action 
in parent and/or hybrid combinations.
2.7.3 The number of effective factors
The most widely used method for estimating the number of effective factors, is the 
Castle-Wright formula (Castle, 1921; Wright, 1968), which utilises information on the 
phenotypic means and variances of the parent lines and subsequent generations (Fi, 
F2, Bi, B2 etc)
The Castle-Wright formula is based on the assumptions that,
i. all segregating loci contribute equally to the trait,
ii. no linkage exists among loci affecting the trait,
iii. no dominance, or the degree of direction of dominance of plus factors is similar
for all loci,
iv. all plus factors are contributed by one parent and all minus factors by the other,
v. no epistasis occurs among alleles at contributing loci, and
30
vi. environmental and genotypic variances are independent and combine additively 
to give total variability.
Since Castle (1921) and Wright (1968) developed their formula for inbred lines, a 
number of models have been proposed for the estimation of effective factors in 
genetically variable populations which are mainly modifications of the model given by 
Castle and Wright (Lawrence and Frey, 1976; Lande, 1981; Cockerham, 1986; Zeng 
et al., 1990; Zeng, 1992; Fenster and Ritland, 1994).
Fenster and Ritland (1994), suggested that since the degree of effective dominance 
can change among segregating generations due to epistasis, assumption iii) would 
not be valid and proposed a formula for estimating effective factors which 
incorporated dominance based on the segregating generations.
Lawrence and Frey (1976) also pointed out that if assumption iv) for the Castle- 
Wright formula is not met, then the numerator would be small and the number of 
factors would be underestimated. The range of the F-, segregates instead of the 
parental range was suggested as the numerator in the formula for estimating 
effective factors. They further argued that the range of the segregating population 
was a better estimate than the parental range if the parents did not represent 
genotypic extremes. The range of the segregating population is also appropriate 
when there is non-additive variance involved in the inheritance of the trait (Dixon et 
al., 1991; Lawrence and Frey, 1976; Fenster and Ritland, 1994).
Lande (1981) also gave a modification of the Castle-Wright formula for use with 
genetically variable populations. Cockerham (1986) and Zeng (1992) separately 
proposed methods for correcting sampling error. Cockerham proposed using the 
experimental variance as a correction factor for the squared difference between the 
mean of the two parents, while Zeng (1992) proposed using least squares to 
estimate the genetic variance and a modified estimator.
31
Multize and Barker (1985), showed that the estimated number of effective factors 
increased as heritability and the probability of type I error increased. There was, 
however, a downward bias of the estimated number of effective factors when 
heritability decreased, the probability of type I error lowered and when there was an 
increase in the type II error.
Zeng et al. (1990), observed that the Wright method is of little value in estimating the 
number of loci affecting a quantitative trait since linkage would prevent the estimation 
from exceeding the number of chromosomes. The sample variance may prevent one 
from concluding that a large number of loci are involved. They, therefore, proposed 
that the alternative is to estimate the number of genes using molecular markers.
2.8 Molecular marker technology and its application in cassava breeding 
and genetics
Molecular markers, which include biochemical (isozymes and storage proteins) and 
DNA markers, exist in every genotype. They occur in large numbers and their 
expression is independent of phenotypic value and as such are powerful tools of 
genetic research (Soller and Beckmann, 1983; Beckmann and Soller, 1986). They 
allow geneticists and plant breeders to locate and follow the numerous interacting 
genes that determine a complex trait as well as tagging those controlled by single 
genes. Molecular markers offer many advantages over conventional phenotypic 
markers. They are i) developmental^ stable, ii) detectable in all tissues, iii) 
unaffected by environmental conditions, and iv) virtually insensitive to epistatic or 
pleiotropic effects. They provide a choice of codominant markers which can be 
identified in heterozygotes or dominant markers which are identified as present or 
absent subclasses (Botstein et al., 1980; Williams et al., 1990).
Molecular markers have made an immense contribution to cassava breeding and 
genetics. The areas covered include, the development of genetic maps (Fregene et
32
al., 1997), the assessment of genetic diversity (Bonierbale et al., 1997b; Lefevre and 
Charrier, 1993; Beeching et al., 1993; Mignouna and Dixon 1997; Sanchez et al., 
1999; Fregene et al., 2000), taxonomical studies (Second et al., 1997), 
understanding the phylogenetic relationships in the genus (Calvalho et al., 1993; Roa 
et al., 1997; Oslen and Schaal, 1999; Roa et al., 2000), confirmation of ploidy 
(Lefevre and Charrier, 1993; Fregene et al., 1994) and cultivar identification (Ocampo 
et al., 1993; Wanyera, 1993; Laminski et al., 1997).
2.8.1 Isozyme markers
Isoenzymes or isozymes of a gene product refer to any two distinguishable proteins 
that catalyse the biochemical reaction of the gene. Their usefulness as molecular 
markers depends on the rarity of the allozyme variant (group of isozymes that 
catalyses the same gene product), how tightly linked they are to a gene and how 
difficult a direct screening method is. The main limitation of isozyme markers is that 
only few gene products can be revealed. Nevertheless, isozymes have been used in 
cassava breeding and genetics (Wanyera, 1993; Lefevre and Charrier, 1993; 
Ocampo et al., 1993; Laminski et al., 1997).
Wanyera, (1993) demonstrated the usefulness of isoenzymes in confirming true 
hybrids in a cross between M. glaziovil and M. esculenta. Lefevre and Charrier, 
(1993) detected genetic diversity among several cassava clones with isozyme 
markers. The study also confirmed that cassava is a true diploid based on the 
inheritance of the markers. Ocampo et al. (1993) used the esterase isozyme to 
fingerprint the cassava germplasm collection held at the CIAT. This isozyme was 
also used in fingerprinting South African elite cultivars (Laminski et al., 1997). 
Fregene et al. (1997) placed three Isozyme markers on the cassava genetic linkage 
map developed at CIAT. Isozymes markers were also used to develop a procedure 
for identifying varieties of cassava (Hussain et al, 1987; Ramirez et al., 1987).
33
2.8.2 DNA markers
A DNA marker is basically a small region of DNA showing sequence polymorphism in 
different individuals within a species (Liu, 1997). The markers are based on the 
enormous variation or polymorphism in DNA sequences of organisms. DNA markers 
eliminate the limitations in genome investigations using morphological and isozyme 
markers, such as gene expression and environmental interaction, heritability, and low 
map resolution (Vogel et al., 1996). DNA-based marker systems can also be used for 
indirect selection of tagged loci affecting qualitative or quantitative traits and to 
monitor loci during introgression or selection programmes, thus reducing the number 
of backcross generations (Baird et al., 1996). Many DNA fingerprinting techniques 
have been developed to generate markers and are generally based on one of two 
strategies; classical, hybridisation-based fingerprinting and PCR-based fingerprinting.
Classical, hybridisation-based fingerprinting which was first developed for use in 
human genetics and is now used as a general method of genetic analysis involves 
the cutting of genomic DNA with restriction enzymes followed by electrophoretic 
separation of the DNA fragments. The restriction fragment length polymorphism 
(RFLP) generated is then detected with probes targeted to specific regions of the 
genome, after southern blot transfer to a nylon membrane (Soller and Beckman, 
1983 ).
PCR-based fingerprinting involves the amplification of particular DNA sequences 
using specific or arbitrary primers and a thermostable polymerase. Amplification 
products are separated by electrophoresis and detected by staining with ethidium 
bromide on agarose gels, silver on polyacrylamide gels or the use of radioactive or 
fluorescent labelled primers in the amplification reaction (Ehrlich et al., 1989). 
Techniques in this category include random amplified polymorphic DNA (RAPD) 
(Williams et al., 1990), DNA amplification fingerprinting (DAF) (Caetano-Anolles et
34
a/., 1991), arbitrarily primed-PCR (AP-PCR) (Welsh and McClelland, 1990) 
microsatellites (Hamada et al., 1982) and amplified fragment length polymorphism 
(AFLP) (Vos etal., 1995).
DNA markers vary in their level of polymorphism and the amount of information 
generated. The most informative DNA markers are characterised by high 
polymorphism information coefficient (PIC), indicating relatively large number of 
alleles with similar frequencies in each locus (Botstein et al., 1980). Polymorphism 
may also be due to single site alterations due to mutations, which abolish or create a 
restriction site, insertions, deletions, or inversions between two restriction sites. In 
this case, the level of polymorphism is low. Polymorphism may also be due to 
variable number of tandem repeats (VNTRs) resulting in a marker system with high 
levels of polymorphism and can detect variation between closely related relatives. 
Markers that detect single site alterations include RFLPs, RAPD, and AFLPs, while 
minisatellites and microsatellite markers detect VNTRs.
2.8.3 Restriction Fragment Length Polymorphism (RFLP)
RFLP is a powerful tool in genetic mapping. RFLP markers are cloned DNA 
segments that can be used to reveal base pair changes or rearrangements in 
homologous DNA sequences (Beckman and Soller, 1986). They are codominant 
markers and scoring is simple because of the low copy number. However, the use of 
RFLP markers requires that many probes are used which requires a complementary 
DNA (cDNA) or genomic library of the species already exist.
The use of RFLP markers has led to the rapid construction of linkage maps in many 
plant species and the placement of genes that control qualitative traits on these 
maps. Single locus disease resistance genes were the first category of genes to be 
mapped extensively using RFLPs. This was partly because most characterised 
disease resistance genes are controlled by alleles at a single locus and most are
35
dominant genes that are easy to assay. RFLPs have also been used to study and 
identify markers that are tightly linked to genes controlling quantitative traits 
(Tanksley, 1993; Young, 1996). In RFLP mapping, regions of interest are detected 
based on their map positions by analying flanking markers when the defined intervals 
are small.
RFLP markers have contributed to DNA marker technology in cassava. Angel et al. 
(1993) initiated work on a detailed genetic map of cassava for tagging agronomically 
important traits and to clone cassava genes. Currently the map has 239 RFLP 
markers (Fregene et al., 2001). Beeching et al., (1993), assessed the genetic 
diversity of several accessions of cassava with RFLPs. Recently, Oslen and Schaal 
(1999) demonstrated the geographic origin of cassava from its relative, M. esculenta 
subspecies flabellifolia based on sequence variation of the single copy nuclear gene 
glyceraldehyde 3-phosphate dehydrogenase (G3pdh).
Despite the potentials of RFLP markers in breeding, there are limitations. It is 
laborious and intensive and does not generally offer the opportunity to target tight 
linkages to specific loci because of the specific probe used. In addition, RFLPs 
cannot detect all the differences that exist between nucleotides, only those that occur 
at restriction sites. The value of the marker comprising an RFLP map depends on the 
degree to which such markers can be used in a population other than the mapping 
population. This is a function of the degree of genotypic diversity in the primary gene 
pools (i.e the mapping population) but also depends on the usefulness of the probes 
used.
Combinations of RFLP and RAPD markers are also commonly used in the 
constructions of linkage maps for many plant species (Fregene et al., 1994; Zhang et 
al., 1996). Calvalho et al. (1993) used RFLP of mtDNA and RAPDDs to identify 
phylogenetic relationships in the Manihot genus.
36
2.8.4 Random Amplified Polymorphic DNA (RAPD)
RAPD primers usually have 10 base pair sequences that are used to amplify 
unknown regions of a genome. The short sequences of the primers make possible a 
multitude of primer binding sites throughout the genome. Efficient amplification of 
DNA fragments may occur when two primer binding sites occur in close proximity 
(Burrow and Blake, 1998). RAPD markers are dominant since the polymorphism is 
detected as a failure of one allele to amplify due to mutations in the primer-binding
RAPD markers are the most extensively used markers in cassava biotechnology, 
especially in determination of phylogenetic relationships and genetic diversity in 
Manihot species (Marmey et al., 1994; Laminski et al., 1997; Schaal et al., 1997). 
They are simple, rapid, and only small amounts of DNA are required. Gomez et al. 
(1996) subjected 328 RAPD markers to linkage analysis in cassava. Out of this, 30 
were placed on the cassava linkage map at CIAT (Fregene et al., 1997). Currently, 
the map has 80 RAPD markers (Fregene et al., 2001). Marmey et al. (1994) 
demonstrated genetic diversity among African cassava accessions using RAPD 
markers. Mignouna and Dixon (1997) observed that several African landraces with 
varying levels of resistance to CMD were genetically different from the genetic stock 
58308 based on RAPD analysis.
DAFs and AP-PCR are basically variations of the RAPD technique. DAF differ from 
RAPDs, in that the primers are made up of short (sequences 5-8 nucleotides long) 
and the fragments are separated on polyacrylamide gels to give better resolution of 
the DNA fragments. DAFs and AP-PCRs have not been reported in cassava.
2.8.5 Variable number of tandem repeats (VNTR)
The two classes of VNTR markers are the minisatellite with repeat units from about 
10-45 base pairs and microsatellite or simple sequence repeats (SSR) with repeat
37
units of 2-4 nucleotides (Jeffrey et al., 1985). Eukaryotic genomes contain extended 
repetitive elements, many of which are mobile such as the retrotransposons which 
are flanked by long terminal direct repeats (LTRs) (Rhode, 1996). The presence or 
absence of flanking LTRs which occur in high copy numbers in the genome have 
been characterised in a number of plant species and have been the basis to 
establish a PCR-based marker system known as the inverse sequence-tagged 
repeats (ISTR).
2.8.6 Microsatellite or simple sequence repeats (SSRs)
Microsatellites or simple sequence repeats (SSR)s are well distributed within the 
genome of eukayotes (Morgante and Olivieri, 1993). They are flanked by unique 
sequences that serve as templates for specific primers to amplify the SSR allele via 
the polymerase chain reaction. Microsatellite polymorphism which can be detected 
by autoradiography, fluorescence labelling, silver staining and ethidium bromide 
staining have been used to complement both the RFLP and PCR based markers.
SSRs have a high level of allelic diversity as a result of the variable number of repeat 
units within their structure, and this makes them valuable as genetic markers 
(Hamada et al., 1982; Morgante and Olivieri, 1993). They are often multiallelic, can 
be multiplexed, automated for high throughput genotyping and convenient to 
exchange between laboratories (Powell et al., 1996b; Chen et al., 1997). They are 
especially suited for distinguishing between closely related genotypes. Although the 
procedure for obtaining microsatellites is laborious and expensive, their conversion to 
PCR markers allows the screening of large numbers of alleles for defined loci. The 
markers are codominant and are, therefore, a suitable option for mapping and 
molecular characterisation in cassava (Agyare-Tabbi et al., 1997; Chavarriaga- 
Aguirre et al., 1998).
38
Chavarriaga-Aguirre et al., (1998) isolated and characterised 14 highly heterozygous 
GA rich microsatellite DNA in cassava. A total of 521 accessions from the cassava 
core collection at CIAT were successfully screened with four SSR loci developed into 
PCR-based markers, to determine genetic diversity (Chavarriaga-Aguirre et al., 
1999). Cassava microsatellites were also used to assess genetic diversity among 
cassava accessions and assess genetic relationships between cassava and its wild 
relatives (Roa et al., 2000). Recently, Mba et al. (2001), developed and characterised 
172 new SSR markers and placed 36 of these on the cassava linkage map.
2.8.7 Amplified fragment length polymorphism (AFLP)
The amplified fragment length polymorphism (AFLP) technology, which is based on a 
combination of restriction digestion and PCR, was developed by Vos et al. (1995). It 
involves the selective amplification by PCR of a subset of restriction fragments of 
agenomic sample. DNA is digested with two restriction enzymes, a frequent cutter 
and a rare cutter. Double-stranded DNA adapters are then ligated to the ends of the 
DNA fragments to generate template DNA for amplification. Thus, the sequence of 
the adapters and the adjacent restriction site serve as primer binding sites for 
subsequent amplification of the restriction fragments by PCR. Selective nucleotides 
extending into the restriction fragments are added to the 3' ends of the PCR primers 
such that only subsets of the restriction fragments are recognised. Only restriction 
fragments in which the nucleotides flanking the restriction site match the selective 
nucleotides will be amplified. The subset of amplified fragments are then analysed by 
denaturing polyacrylamide gel electrophoresis to generate the fingerprint.
The frequent cutter generates fragments with the optimal size range for separation 
on denaturing gels and the number of fragments to be amplified is reduced by using 
the rare cutter, since only the rare cutter/frequent cutter fragments are amplified (Vos 
et al., 1995). In most AFLP studies, genomic DNA is digested with the enzymes
39
EcoR1/Mse1 that is suitable for species with AT (adenine and thiamine) rich 
genomes (Vos et al., 1995; Janessen et al., 1996). Other enzyme combinations such 
as Pst1/Mse1, Hindll/Mse1 are useful for genomes with 40-50% GC (guanine and 
cytocine) content, while Apa1/Taq1 are used for GC rich genomes (Wong et al.,
1999).
AFLP markers have high multiplex ratios since several genomic fragments are 
analysed in on assay and require no prior sequence data (Rusell et al., 1997; Powell 
et al., 1996a). They are generally dominant markers based on the presence or 
absence of a particular product. However, with appropriate software, such as 
Keygene (Wageningen, the Netherlands) they can be scored as codominant markers
The main advantage of AFLP technology over other marker systems is that variability 
can be assessed at a large number of independent loci and data can be obtained 
quickly and reproducibly (Sanchez et al., 1999). AFLP markers have been 
extensively used in cassava molecular marker technology (Bonierbale et al., 1997b; 
Second et al., 1997; Roa et al., 1997; Sanchez et al., 1999; Fregene et al., 2000). 
Fregene et al. (2000) were able to detect a considerable number of duplicates in 
some African landraces with AFLP markers. They also detected a unique AFLP 
fragment among the African landraces, which was associated with branching pattern. 
Sanchez et al. (1999) established genetic similarities among cassava accessions 
resistant to the CBB disease. AFLP markers were also used to demonstrate genetic 
relationships among different cassava accessions and other manihot species 
(Bonierbale et al., 1997b; Second et al., 1997; Roa et al., 1997).
2.9 Determination of molecular markers associated with traits of interest
The strategies available to facilitate detection of molecular markers associated with 
specific traits of interest include bulk segregant analysis (Michelmore et al., 1991; 
Giovannoni et al., 1991), linkage mapping (Tanksley et al., 1989) and QTL analysis
40
(Dudley, 1993; Tanksley, 1993). For markers associated with disease resistance 
genes, degenerate primers based on isolated candidate resistant genes can be used 
to screen for resistance gene analogs (Kanazin et al., 1996; Collins et al., 1998).
2.9.1 Bulk segregant analysis
Bulk segregant analysis (BSA) is a rapid method for identification of markers in 
specific regions of the genome (Michelmore et al., 1991; Giovannoni et al., 1991). 
The method is especially useful when there are no other markers available within the 
region of interest. It involves comparing two pooled DNA samples of individuals from 
a segregating population. The two DNA pools contrast for a trait (e.g. resistant and 
susceptible) or a marker allele in a previously mapped population are analysed to 
identify markers distinguishing them. The individuals in each bulk are identical for the 
trait or gene of interest but are arbitrary for other traits. So the genomic region is 
studied against a randomised genetic background of unlinked loci. Markers that are 
polymorphic between the pools will be genetically linked to the loci determining the 
trait used to make up the pools.
Giovannoni et al. (1991) demonstrated the use of DNA pooling to saturate specific 
chromosomal regions with additional markers, based on DNA pools from an existing 
tomato mapping population. Kiehne and Neale (1998) adapted this strategy in the 
outcrossing species loblolly pine, and identified nine new RAPD markers linked to a 
chromosomal interval containing the previously mapped QTL for wood specific 
gravity. Reamon-Buttner et al. (1998) detected nine AFLP makers, tightly linked to 
the sex locus of the dioecious plant Asparagus officinalis L, which is important to 
plant breeding in developing sex-specific PCR primers.
Bulk segregant analysis has also been explored in tagging molecular markers linked 
to major disease and pest resistance genes (Michelmore et al., 1991; Haley et al., 
1993; Miklas et al., 1993; Garcia et al., 1996; Seo et al., 1997; Fritz et al., 1999;
41
Correa et al., 2000). Michelmore et al., (1991), were the first to report three RAPD 
markers linked to major disease resistance genes using contrasting DNA bulks 
composed of F2 individuals of known genotype. Akano et al. (2002) recently used 
BSA to identify a SSR marker linked to the CMD resistance gene, CMD2. Molecular 
markers linked to two major rust resistance genes in common bean, were detected 
using introgression and bulk segregant analysis (Miklas et al., 1993; Haley et al.,
1993). The bean studies, however, revealed that a major limiting factor in the 
application of dominant markers linked in coupling with dominant alleles or in 
repulsion with recessive resistant alleles is the amplification of the DNA fragments of 
equal sizes in susceptible germplasm (Miklas et al., 1996). Sequencing results of a 
RAPD marker, which distinguished between resistance and susceptibility to the 
Russian wheat aphid (Fritz et al., 1999) revealed a deletion in the susceptible parent 
relative to the resistant parent. Markers of interest are therefore, usually developed 
into sequence characterised amplified regions (SCARs) which are codominant 
markers for resistance screening (Garcia et al., 1996; Seo et al., 1997; Fritz et al., 
1999; Correa et al., 2000).
2.9.2 Linkage mapping
Linkage maps are based on recombination frequencies. The percentage of a 
segregating progeny that are recombinants for a pair of linked loci is the 
recombination frequency. The recombination frequency gives an estimate of the 
distance between two loci in a chromosome, on the assumption that the probability of 
crossover is proportional to the distance between loci.
Construction of genetic linkage maps requires that the most appropriate mapping 
population(s) is available. Pairwise recombination frequencies of these populations 
are calculated from genetic data such as DNA marker scores, linkage groups are 
established and map distances estimated, and then map order is determined (Staub
42
et al., 1996). Different marker systems may be included in the analysis of the 
population. Computer packages such as Linkage 1 (Suiter et al., 1983), GMendel 
(Holloway and Knapp, 1994), Mapmaker (Lander and Botstein, 1986; Lander et al., 
1987), MapManager (Manly and Cudmore, 1996) and JoinMap (Stam, 1993) have 
been developed to aid in the analysis of genetic data for map construction. These 
programmes use data obtained from segregating populations to estimate 
recombination frequencies that are then used to determine the linear arrangement of 
genetic markers by minimising recombination events.
Genetic linkage maps can provide a more direct method for selecting desirable 
genes via their linkage to easily detectable molecular markers (Tanksley et al., 1989). 
Once a trait is identified and mapped, marker-assisted selection (MAS) could be 
used in early identification of the trait. This would reduce, breeding population sizes, 
continuous recurrent testing and the time required in developing a superior line. The 
cassava genetic linkage map developed at CIAT (Fregene et al., 1997), currently has 
five known genes including two CMD resistance genes CMD1 and CMD2 placed on it 
and a number of quantitative trait loci (QTL) associated with some linkage groups 
(Fregene et al., 2001).
2.9.3 Quantitative trait loci (QTL)
Molecular markers are being utilised in studying quantitative traits loci (QTLs) that 
affect traits and have been suggested as tools to enhance the efficiency of selecting 
quantitative traits in breeding programmes (Dudley, 1993; Young, 1996). The theory 
of QTL mapping was first described in 1923 by Sax and later elaborated by Thoday 
in 1961 who suggested that if the segregation is simply inherited, monogenes could 
be used to detect linked QTLs (Thoday, 1961). Young (1996) suggested that defined 
sequences of DNA, in modern day QTL mapping act as the linked monogenic 
markers used to detect QTLs.
43
Mapping QTLs with molecular markers require the analysis of large segregating 
populations (Tanksley, 1993). A population derived from distinctively different parents 
(e.g. resistant and susceptible) is analysed with the marker to test for linkage of 
markers to a QTL. This is based on its position on the map from flanking markers. 
The population is then classified into homozygous or heterozygous for the marker of 
the parents and compared statistically with phenotypic values to determine linkage of 
the molecular marker with the QTL. One of the first published reports of QTL 
mapping with DNA markers involved fruit size, pH and soluble solids in tomato 
(Paterson etal., 1988).
QTL mapping for quantitative disease resistance has been reported in a number of 
plant species (Young, 1996). Wang et al. 1994, mapped QTLs affecting resistance to 
the rice blast fungus in a population consisting of recombinant inbred lines. QTL’s 
associated with resistance to potato late blight was studied in a Ft cross between two 
highly heterozygous parents where the phenotypic disease responses were 
compared with DNA marker data (Leonards-Schippers et al., 1994). The results 
indicated that resistance to the pathogen is a cumulative phenotype resulting from a 
small number of race-specific and race-nonspecific resistance loci. Other QTL 
mapping studies on quantitative disease resistance include resistance to leaf spot 
fungus of maize, resistance to the soybean cyst nematode, and tomato bacterial wilt 
(Bubeck et al., 1993; Danesh et al., 1994; Concibido et al., 1994). Recently, QTLs 
associated with dry matter yield and dry matter percentage have been linked to 
linkage group D of the cassava framework map (Fregene et al., 2001).
A number of computer software have been developed for QTL analysis, these 
include MAPMARKER/QTL, QTLSTAT, LINKAGE, MAPQTL, QGENE, Map Manager 
QT and PGRI (Liu, 1998).
44
2.9.4 Screening for resistance gene analogs
Genes conferring resistance to several diseases caused by a wide range of 
pathogens including viruses, fungi, bacteria and nematode have been isolated using 
map-based cloning or insertional mutation (Michelmore, 1995; Staskawicz et al., 
1995). Most of the cloned resistance genes have been found to have common 
structural motifs known as leucin rich repeats (LRR) which are actually the specificity 
determinants (Thomas et al., 1997; Wang et al., 1998). Leucin rich repeats 
containing resistant genes can be grouped into the transmembrane R proteins which 
lack nucleotide binding sites (NBS) and the cytoplasmically located R proteins with 
NBS (Barker et al., 1997; Ori etal., 1997).
The presence of these conserved sequences among R genes has provided an 
opportunity for isolating new resistant genes or resistant gene analogs (RGAs) based 
on PCR methods by designing degenerate primers based on the conserved 
sequences (Kanazin et al., 1996; Collins et al., 1998; Gentzbittel et al., 1998).
The cloning and analysis of these resistant gene analogs has opened new avenues 
of research on plant disease and produced markers tightly linked to resistance 
genes. Gentzbittel et al., (1998) isolated RGAs in sunflower using NBS and protein 
kinase conserved motiffs. They further mapped three of the clones to linkage group 1 
where downy mildew resistance genes are located. RGA were also mapped to four 
downy mildew R genes clusters in lettuce (Shen et al., 1998).
2.9.5 Sequence tagged sites (STS)
A recent addition to the tools for fine physical mapping of genes of interest is 
sequence tagged sites (STS). An STS is a short DNA sequence generally, between 
100-500 bp that is easily recognisable and occurs only once in a chromosome or 
genome studied (Brown, 1999). STS are derived in the genome from expressed 
sequence tags (EST), which are obtained from cDNA clones; simple, sequence
45
length polymorphism (SSLP), or from random genomic sequences. To date, six ESTs 
have been placed on the cassava genetic map (Fregene et al., 2001). Single 
nucleotide polymorphism (SNPs) are another group of site specific markers, which 
are the most common type of DNA sequence variation and occur every 100 to 300 
base pairs (Lohmann et al., 2000). Due to their high frequencies, the chances are 
high that disease resistance or susceptibility would be caused by or closely 
associated with specific SNPs.
2.9.6 Type of population
The choice of an appropriate mapping population depends on the type of marker 
systems employed (Tanksley et al., 1988). Generally F2s and backcross progenies 
are used in the construction of linkage maps, QTL analysis, and BSA for 
homozygous species. In a heterozygous species, the segregating F-, is used (Liu, 
1998; Pillay and Kenny, 1996; Fregene et al., 1997). However, this should be an 
informative cross between two divergent parents that are heterozygous for the 
markers (Falconer and Mackay, 1997). The cassava genetic map at CIAT is based 
on an Ft population of two geographically divergent parents. The female parent TMS 
I30572, having resistance from clone 58308, while the South American male parent 
has no resistance to CMD (Fregene et al., 1997).
Maximum genetic information is obtained from a completely classified F2 or 
segregating Ft population using a codominant marker system. Information from a 
dominant marker system can be equivalent to a completely classified F2 population if 
progeny tests (i.e. F3 or F2, BC) are used to identify heterozygous F2 individuals 
(Staub et al., 1996). Progeny testing is required when the phenotypes do not 
consistently reflect genotype or where trait expression is controlled by QTLs.
Recombinant inbred lines (RIL) are suitable for QTL mapping in that they consist of 
many distinct lines each a mosaic of different homozygous segments from the
46
original parent. Dominant markers also supply as much information as codominant 
markers in RIL (i.e. an array of genetically related lines; usually F8 or more), doubled 
haploid, or backcross populations in coupling phase because in these populations, all 
loci are homozygous, or nearly so (Burr and Burr, 1991 ;Liu, 1997).
Other types of population used in identification of markers tightly linked to disease 
resistance genes are the near-isogenic lines (NILs) (Muehlbauer et al., 1988; Young 
et al., 1988). A donor-parent (DP) carrying the gene of interest is repeatedly 
backcrossed to a recurrent parent (RP), selection of the desired gene and recovery 
of the RP phenotype is continued until the newly developed line is theoretically nearly 
isogenic with the RP, except for the chromosome segment containing the target 
gene. Linkage between a molecular marker and the target gene can be assessed by 
determining the marker genotype of the RP, its various NIL derivatives, and their 
corresponding DPs (Melchinger, 1990). Although the development of NILs through 
back11 crossing for economically important genes is costly and time consuming 
(Kelly, 1995), the use of such NILs has permitted the successful tagging of three 
disease resistance genes in common bean (Haley et al., 1994; Miklas et al., 1996).
Inter-gene-pool crossing technique was used successfully to tag two rust resistance 
genes in common bean (Haley et al., 1993; Miklas et al., 1993) and in tomato, 
interspecific populations have been applied in mapping of some disease resistance 
genes (Paterson et al., 1990).
47
Chapter Three 
Materials and Methods
The study was carried out in the experimental fields and at the Cellular and Molecular 
Technology Laboratories (CMTL) of the International Institute of Tropical Agriculture 
(1ITA) in Ibadan, Nigeria.
3.1 Analysis of variance and combining ability analyses
3.1.1 Genetic Material
Diverse cassava genotypes with varying levels of resistance and susceptibility to 
CMD, including the resistant genetic stock 58308, six improved clones and 15 African 
landraces and the progenies from their F-) crosses were used in the study. The 
cassava genotypes, TMS I30572 (moderately resistant), TMS 91/02324 (resistant to 
CMD), a resistant landrace TME1 (Antiota) and a susceptible landrace TME117 
(Isunikankiyan) were included as checks. The parental and check genotypes are 
described in Table 3.1.
3.1.2 Experimental design
Seeds of the Ft crosses, parental and check genotypes were provided by the 
cassava breeding unit at IITA for the study. The seeds were obtained from two 
mating designs. One was a 4X4 diallel mating scheme in all possible combinations 
including selfs and reciprocals. It involved the resistant genetic stock 58308 and 
three improved clones, TMS 130001, TMS I30572 (resistant to CMD), and TMS 
I30555 (susceptible to CMD). The other mating design was a 3X18 North Carolina 
design II (NCD II) with TMS 130001, TMS I30572, and TMS I30555 as the female 
parents, three resistant improved clones and the 15 landraces were the male 
parents.
48
Table 3.1 Description of cassava clones used as parents in 4X4 diallel and 3X18 North 
Carolina Design II matings and checks
Clone Pedigree information, local name and origin Genetic Experiment
Diallel Design II
TMS 130001 Pedigree information lost Parent 1 Female 1
TMS I30555 58308 X Oyarugba dudu Parent 2 Female 2
TMS I30572* 58308 XBranca de Santa Caterina (OP)+ Parent 3 Female 3
TMS 160142 KR685 OP Male 1
TMS I90257 58308 X Oyarugba dudu Male 2
TMS 14(2)1425 58308 X Oyarugba fufun Male 3
58308 M. esculenta X M. glaziovii Parent 4
91/02324* TME1 OP
TME1* Antiota (Ondo, Ondo State, Nigeria) Male 4
TME2 Odungbo (Opeji, Ogun State, Nigeria) Male 5
TME4 Atu (Iwo, Kwara State, Nigeria) Male 6
TME5 Bagiwawa (New Busa, Niger State, Nigeria) Male 7
TME6 Lapai-1 (Lapai, Niger State, Nigeria) Male 8
TME7 Oko-lyawo(New Lapai, Niger State, Nigeria) Male 9
TME8 Amala (Ireuekpen, Edo State, Nigeria) Male 10
TME9 Olekanga (Ogbomosho, Oyo State, Nigeria) Male 11
TME10 Orente (Ogbomosho, Oyo State, Nigeria) Male 12
TME11 Igueeba (Warri, Delta State, Nigeria)- Male 13
TME12 Tokunbo (Ibadan, Oyo State, Nigeria) Male 14
TME14 Abbey Ife (Abbey-lfe, Osun State, Nigeria) Male 15
TME31 Bakince-lri (Bahago, Sokoto State, Nigeria) Male 16
TME41 Danbusa (Kanji, Niger State, Nigeria) Male 17
TME117* Isunikankiyan (Ibadan, Oyo State, Nigeria) Male 18
'C lone used as check 
+OP=open pollinated
49
In 1997, the seedlings from these crosses were established in a seedling nursery in 
Mokwa, Niger State, Nigeria, to produce woody cuttings for the experiments. Mokwa 
is in the Southern Guinea Savanna with a ferric luvisol soil type. It is generally a low 
disease pressure site for CMD.
The genetic material was then established in two experiments at Mokwa and Ibadan 
(Oyo State, Nigeria) during the 1998 growing season and in Ibadan during the 1999 
growing season. Ibadan is in the forest-savanna transition with a ferric luvisol soil 
type. The physical and chemical properties of soil and rainfall pattern of the two 
locations during the two growing seasons are described in Appendix I. In Mokwa, the 
two experiments together with their parents, and three checks (TMS I30572, TME1, 
and TME117) were evaluated, while in Ibadan, the two experiments, their parents 
and four checks (91/02324, TMS I30572, TME1, and TME117) were evaluated.
The design for each experiment was a randomised complete block with two 
replications. Plants were established using 20 cm long cuttings. Plants were spaced 
at 0.5 m apart in rows (ridges 30 cm high and 10 m long) which were 1 m apart giving 
a plant population of 20,000 plants per hectare. Due to differential seed set of the 
parents, the number of progenies varied for the different crosses. The number of 
progenies obtained from the seedling nursery ranged from 16 to 300 in the diallel 
experiment and 52 to 934 in the NCD II experiment. To ensure the survival of each F-, 
genotype, two cuttings per genotype were planted in each replicate. The second 
cutting was removed at 6 weeks after planting (WAP) when the plants were 
established. Twenty cuttings of each parental or check genotype were planted in 
each replicate with the same spacing as the crosses. The experiments were 
evaluated for CMD under rain-fed conditions. No fertiliser or herbicide was applied 
during the course of the experiments. Hand weeding was done when necessary.
50
3.1.3 Assessment of CMD
Individual plants in each F-, cross and 10 plants each of the parents and checks were 
assessed for their reaction to CMD under natural infection by whiteflies. Assessment 
was based on their phenotypic expression of severity of symptoms using the 
standard five point scoring scale system for CMD (IITA, 1990). Disease severity 
scores of the whole plants and shoot tips (from the shoot apex to the first fully 
expanded leaf) were assessed at 6, 12 and 20 WAP in 1998 and at 6 and 12 WAP in 
1999. Table 3.2 describes the five point scoring scale and an illustration of the 
disease severity scores on cassava leaves is presented in Plate 3. 1.
3.1.4 Data analyses
Data collected from the three environments where the experiments were conducted 
were analysed using the statistical analysis system (SAS) package (SAS, 1999). An 
environment was defined as the particular location and year that the experiments 
were conducted. Thus, the three environments were Mokwa 1998, Ibadan 1998, and 
Ibadan 1999.
The weighted means for incidence and disease severity of CMD in shoot apices and 
whole plants of each of the checks, parents, and crosses at the different scoring 
dates were determined from the data collected and used for the genetic analysis. 
Incidence was determined as the percentage of diseased plants or percentage of 
plants with diseased shoot tips. The weighted average disease severity score of a 
parent, check, or cross was defined as the summation of the product of the 
frequency count and the value of the disease severity class, divided by the total 
number of plants evaluated, and was determined for each replicate. If the weighted 
average disease severity score of a parent, check, or Fi progeny cross was less than 
or equal to 2, it was classified as resistant, otherwise it was classified as susceptible.
51
Table 3.2 Standard five point scale scoring systems for CMD (IITA, 1990)
Disease Score Symptom Status
1 No obvious symptoms. Highly resistant (HR)
2 Mild chlorotic patterns or mild leaf distortion at 
the base
Resistant (R)
3 Strong mosaic on the entire leaf, distortions of 
leaves
Moderately susceptible (MS)
4 Severe mosaic distortion, reductions of leave 
lamina with about 2/3 of the leaves affected
Susceptible (S)
5 Severe mosaic, severe distortion of leaves, 
stunting of entire plant and about 4/5 of leaves 
affected
Highly susceptible (HS)
5 2
I
Plate 3.1 Cassava leaves, with varying levels of cassava mosaic disease severity 
CMD 1, No obvious symptoms.
CMD 2, Mild chlorotic pattern, mild leaf distortion at the base.
CMD 3, Mosaic on the entire leaf, distortion of leaf 
CMD 4, Severe mosaic distortion, reduction of leaf lamina.
CMD 5, Severe mosaic, severe distortion of leaf and severe reduction of leaf lamina.
5 3
A simple phenotypic correlation analysis performed on the average CMD incidence 
and disease severity scores of all genotypes in the two experiments showed a 
consistent significant linear relationship between incidence and severity at the 
different scoring dates across environments and in the individual environments 
(Appendix II). Genetic analyses of the F-, crosses were based on CMD severity at 12 
WAP and the assessment of genetic relationships among the parents was based on 
all the CMD severity and incidence responses.
The analyses of variance for each experiment were performed on the individual 
environments and a combined analysis of variance was performed with data from the 
three environments. The analyses were based on mixed models with the genotypes 
considered fixed; replicate, environment, and genotype by environment interaction 
(GXE) considered as random effects.
Appropriate F-tests were performed according to the expectations of the means 
squares, based on the procedures described by McIntosh (1983) for mixed model 
analysis. In the individual environment analyses, the genotypic components were 
tested with the pooled error. For the combined analyses, the environment was tested 
with the replicates nested within environments, Env(Rep) mean square, and the 
genotypic components were tested with their respective genotype by environment 
(GXE) interactions. The GXE effects were tested with the pooled error.
3.1.4.1 3X18 North Carolina Design II (NCD II) analysis
The general linear model (GLM) procedure in SAS which uses the method of least 
squares to fit general linear models was used for the analyses of variance (SAS, 
1999). Mean squares were calculated from Type III sums of squares. Genotypes 
were partitioned into the variation due to test genotypes (parents and crosses) and 
checks. Test genotypes were then partitioned into parents and crosses. The parents 
were further partitioned into female, male, resistant and susceptible parents and one
54
degree of freedom orthogonal contrasts, female versus male parents and resistant 
versus susceptible parents were made to test for significant variation between the 
different sets of parents. A one degree of freedom orthogonal contrast, between 
parents and crosses was also used to test the significance of average midparent 
heterosis.
Crosses were further partitioned into variation due to the GCA (additive) effect of 
males, the GCA effect of females and variation due to the SCA (non-additive) effect 
(male and female interaction). The general linear model used for the analysis of the 
crosses in the NCD II for one environment was as follows:
y ijk ~ M " m i ~.f ~mf  ,j+b A. -i- e jJk  (1)
In this model, y jJk was the observed response to CMD; f.i the general mean; mj the 
GCA effect of the ith male; / ;the GCA effect of jth female; mf :j the SCA effect of 
the ijth cross; bk was the effect of the kth replicate; and &ijk the error associated with
each observation (Singh and Chaudhary, 1985).
The analysis of variance for the combined analysis across environments was 
performed as given for an individual environment. The GXE interactions however, 
were further partitioned into their components.
The general linear model assumed for the NCD II population for the combined 
analysis across environments was:
ym= M + mi + f  j+mf ^ l k+ b m +m l lk+fl Jk+mfl iJk+ e ijkl .............(2)
Here, y jjkl was the observed response to CMD across the three environments;/;, 
m, f  j and mf were as described in equation (1) for an individual environment.
55
The /t th effect was the kth environment. The bl(k) th effect was the kth environment 
nested in Ith replicate. The mljk th effect was the ith male in the kth environment; and 
fl jk was the effect of the jth female in the kth environment. The fin ljjk th effect was 
the ijth cross in the kth environment and eiJkl the residual associated with each 
observation (Beil and Aitkins, 1967).
Combining ability estimates were determined for individual environments and across 
environments using least square means.
GCA of a clone was estimated as
HereXij. was an F-, mean, X i. was the mean of crosses of the ith clone; X .; was
the mean of crosses of the jth clone and X..is the mean of all the Fi crosses (Beil 
and Aitkins, 1967).
The standard errors of these effects were calculated using the methods described by 
Singh and Chaudhary (1985) for a single environment and Cox and Frey (1984) for 
the combined environment.
The SE for GCA was
g, = X ,.-X .. (3)
and the SCA of a cross was estimated as
= X,j . - X i. - X . j  + X.. (4)
M SE(m -l)
rfm
for males in a single environment or, (5)
56
SE(.(.a = J MSE( f  ^ for females in a single environment or, (6)
rfrn
SE,.,., = ^ for males in the combined environment .... (7)
UA'" 11 mfrl
or,
SE = for females in the combined environment (8)
(:CAr 11 mfrl
The SE for SCA was
SE SCA= J ^ E (m  \){f—1) for a sjng|e environment..............(9)
or,
SEsca= I    for the combined environment (10).
mfr(l)
In equations (5) to (10), MSE was the mean square error of the crosses, m and /  
were the number of males or females, r the number of replicates and /-(/) the 
number of replicates and environments. MSwl and M S were the mean squares of 
males or females by environment interaction and MSmP was the mean square of the 
male by female by environment interaction.
The GCA and SCA effects were tested for significance in two tailed tests as
described by Cox and Frey (1984) and the LSD was determined as,
LSDGCA=t{d[fi^ S E 0CA and ................................................. (11)
LSDSC/l=t(iif 005jxSES(:A (Singh and Chaudhary, 1985)................(12).
57
Negative and significant effects where considered as contributing to resistance and 
positive and significant effects were considered as contributing to susceptibility.
3.1.4.2 4X4 Diallel analysis
Genotypes were partitioned into the variation due to test genotypes (parents and 
crosses) and checks using the GLM procedure in SAS (SAS, 1999). Test genotypes 
were further partitioned into parents and crosses. A one degree of freedom 
orthogonal contrast between parents and crosses was used to test for the presence 
of significant average midparent heterosis.
The genetic analysis of the crosses was based on Griffing's method 1, in which the 
analysis is based on data from a complete diallel (crosses selfs and reciprocals) and 
using model 1, which is valid when the genotypes are fixed (Griffing, 1956b). 
Reciprocal effects were further partitioned into the maternal and non-maternal effects 
(Cockerham and Weir, 1977). The analysis was performed on individual 
environments and a combined analysis over all environments using the diallel-SAS 
programmes written by Kang (1994) for a single environment and Zhang and Kang 
(1997) for the combined environment.
The general linear model for an individual environment was;
Y iJk= M+g,+gJ+Sj+r u+bk+eV[..............................................(13)
In this model, Ym was the observed response to CMD; (i was the general mean; g i 
the general combining ability of the ith parent; g . the general combining ability of 
the jth parent; su the specific combining ability associated with the ijth cross; r r the 
reciprocal effect associated with the ijth cross. The bA.th effect was the effect of the 
kth replicate and ejjk the residual effect associated with each observation (Kang,
58
1994). The reciprocal effect r ;/was further partitioned into the maternal effect of the 
ith parent m.t , the maternal effect of the jth parent nij and the non-maternal effect 
ntj of the ijth cross (Cockerham and Weir, 1977).
The general linear model for the combined analysis was;
YijkT M + g i+ g j+ s ij+ r ij+ lk+b,(k)+glH+gljk+slijk+ r lijk+ e ijki........... (14)
In this model; Y,jkl was the observed response to CMD across the three 
environments, ju, g i , g . , s tj and r 0 and its partitions mn  mj and were as 
described in equation (13) for the individual environment analysis. The effect l klh 
effect was the kth environment; b,^lh effect of the Ith replicate nested in the kth 
environment. The g likth effect was the general combining ability of the ith parent in 
the kth environment; the g ljk th effect was the general combining ability of the jth 
parent in the kth environment and the s ljk{h effect was the specific combining ability 
associated with the ijth cross in the kth environment. The r lijk th effect was the 
reciprocal effect associated with the ijth cross in the kth environment and eijkl the
residual effect associated with each observation (Kang et al., 1999). The interaction 
of the reciprocal effects with the environment was also further partitioned into the 
maternal by environment interaction (mllk and ml jk) and non-maternal by
environment (nl.jk) interaction .
Estimates of GCA, SCA, reciprocal maternal and non-maternal effects of the crosses 
were generated by the SAS programme with least square means. The computational 
formulae, which generated the estimates, were based on Cockerham’s statistical 
model (1963) for Griffing's Methods, which was reviewed by Zhang and Kang (1997).
59
3.1.4.3 Relative importance of general and specific combing ability
The relative importance of GCA and SCA in predicting progeny performance can be 
determined from the ratio of the additive variance to total genetic variance. Since the 
genotypes were considered fixed, estimation of variance components was not 
appropriate. The analogous mean square components of the fixed effects were 
therefore estimated to compute the ratios (Baker, 1978; Kang, 1994). The mean 
square components of the fixed GCA effects of males, females, and the SCA effect 
of the female by male interaction were estimated from the variance components of 
the random effects associated with these fixed effects using the MIVQUEO option of 
the VARCOMP procedure in SAS.
For the NCD II the ratio was estimated from the relationship;
.......................................................... (15,
GCA + SCA
Here, GCA was the sum of the GCA mean square components due to female and 
GCA mean square components due to male. SCA was the mean square component 
due to the female by male interaction (Cukador-Olemedo et al., 1997).
For the diallel analyses, the ratio was estimated from the relationship
2 GCA
.(16)
2 GCA + SCA
where GCA and SCA are the mean square components of the GCA and SCA effects 
respectively (Kang, 1994). Mean square component of the GCA and SCA effects in 
the diallel were also estimated from the expected mean squares as described by 
Singh and Chaudhary (1985). The GCA component (# . ..) and SCA component
{6srA) in one environment were estimated as described by Singh and Chaudhary 
(1985) as follows;
60
a 1 V ' „ 2 . . MSGCA MSe
(!CA 2 r
and
 ( 1 8 >
Where, MS GCA,MS SCA and MSs were the mean square for GCA, SCA and Error and 
r was the number of replicates.
For the combined analysis, the GCA component (daCA) and SCA component {6SCA) 
were estimated as described by Zhang and Kang (1997) as follows:
a -  1 v  „ 2 _ MSgca MSGCAi
~------- 2re.........    <19)
and
a  -  2  V  V  -  2 -  M S s c j ~ M S s c m b
’“ ' W 1y )L L  — ™  ......................
Where, MS GCA,MS SCA, MSe and r were as described for an individual environment, 
MSqcaxe, MSscaxe and e were the mean square for GCA X E, SCA X E and number 
of environments respectively.
Pearson correlation was performed on the parent means and GCA effects to compare 
per se performance of the parents with their GCA effects and their topcross 
performance in the NCD II experiment.
61
3.1.4.4 Heritability
Heritability across environments and the exact confidence intervals were estimated 
using the methods described by Knapp et al., (1985) based on progeny mean. 
Heritabilty across environments for each experiment was therefore estimated as,
where MScmss and MSCXE were the mean squares of the crosses and crosses by 
environment across environments respectively. The exact 90% (1-a =0.90) 
confidence limits were determined as
where F 0 0Sutrd/. and F %x.df„df were the tabulated values of F at 0.05 and 0.95 for 
the appropriate degree of freedom in each experiment, MScrmx and MScxr were as 
described above.
3.1.4.5 Heterosis
Least square means were used to estimate average heterosis in each experiment 
and in the individual crosses. Heterosis of a cross was determined as the difference 
between the mean of a cross and the mean of the two parents (mid-parent), or the 
mean of the better parent (high parent), expressed as a percentage of the parent 
mean. Significance of average mid-parent heterosis was estimated by a one degree 
of freedom orthogonal contrast (parent versus cross) from the analysis of variance of 
the test genotypes. Significance of heterosis in an individual cross was determined
(21)
1 -  [ ^ o . o (MS )]_1 ■ uPPer limit (22)
1 - [ ^ ^ / - ^ r j o w e r N m it
CXE
(23)
62
using a least significant difference (LSD) between the parent and cross mean (Dixon 
et al., 1990), given below
where, t(df, 0.05) was the tabular value of t at the 0.05 level of significance with the 
appropriate degrees of freedom, rwas the number of replicates and environments, a 
and b were the number of experimental units in a cross and parental mean 
respectively and MSe was the pooled error mean square. Since the average disease 
severity scores per replicate were used for the GLM analysis to generate the least 
square means used in the estimation of heterosis, the number of experimental units 
in the Fi or parent means was 2.
3.2 Genetic relationships among the various sources of resistance to CMD
The genetic relationship between the different sources of resistance was assessed 
based on the phenotypic similarities of the parental responses to CMD and the 
segregation patterns of the Fi progenies with respect to their response to CMD. The 
experimental procedure and CMD assessment were as described in section 3.1 
above.
3.2.1 Area under the disease progress curve (AUDPC) of the sources of 
resistance to CMD
The mean CMD severity scores in the shoot tip and whole plant at 6 and 12 or 20 
WAP of the parental and check clones which represent the sources of resistance 
were used to estimate the area under their disease progress curve (AUDPC) values 
in each environment and across environments. AUDPC was calculated using the 
method described by Jeger and Viljanen-Rollinson (2001) which allows estimates of 
AUDPC with values from two assessments in the equation,
(24)
63
AUDPC = T + (25)
r
Here, y0 was the CMD value at 6 WAP, and T was 12 or 20 weeks. For Ibadan 1999 
and Mokwa 1998 environments, yTwas the CMD value at 12 WAP while for Ibadan 
1998 and across environments it was 20 WAP and rthe rate.
The rate r, was determined as
AUDPC values were then subjected to Proc Rank procedure in SAS (SAS, 1999) to 
assign the lowest rank value to the least (most resistant) AUDPC values. Spearmans 
correlation (SAS, 1999) was performed on the assigned ranks to test the association 
of the AUDPC values across environments.
3.2.2 Principal component analysis (PCA) of sources of resistance to CMD
Principal component analysis (PCA) was performed on the mean CMD incidence and 
severity scores of the various sources of resistance for the different scoring dates in 
the three environments, using the correlation matrix of the means (SAS, 1999). A 
scatterplot of the PCA scores of the first and second principal components was 
generated to explore the genetic similarities among the parental clones. The aim was 
to group the clones according to their CMD responses.
ln ( -^ M - ln ( - ------
i - y r i - y o 
T
(26).
64
3.2.3 Gene complementarity among various sources of resistance to CMD 
based on the distribution of the Fi disease severity scores
3.2.3.1 Transgressive segregation for resistance to CMD
The mean disease severity scores of individual Fi progenies of the various sources 
of resistance at 12 WAP (largest variation for CMD severity) in an environment, were 
used to construct a frequency table of the distribution of disease severity scores in 
the three environments. The percentage of positive transgressive segregants in the 
Ft cross was determined from the distribution of disease severity scores. A positive 
transgressive segregant was defined as an F-, progeny at least one disease severity 
score lower than the average disease severity score of the better parent in a cross. 
Positive transgressive segregants were determined for each cross in each 
environment and across environments.
3.2.3.2 Cochran-Mantel-Haesnzel
The Cochran-Mantel-Haesnzel ANOVA statistic (SAS, 1999), was used to determine 
significant differences in the mean distribution of the Ft disease severity scores 
among the various sources of resistance and to test for reciprocal differences.
3.2.3.3 Number of effective factors (NE) contributing to resistance to CMD
The minimum number of independently segregating effective factor pairs responsible 
for resistance to CMD was estimated in the segregating Ft populations using a 
modification of the Castle-Wright formula (Castle, 1922; Wright, 1968),
where NE was the number of effective factors by which the parents in a cross differ; 
P-i and P2 were the mean values of the better parent and poorer parent respectively,
(27)
65
VarF2 , the variance of the segregating population and VarE the environmental 
variance estimated from the uniform population. The assumptions for an unbiased 
estimate using this formula are that:
i. all segregating loci contribute equally to the trait,
ii. no linkage exist among loci affecting the trait,
iii. no dominance, or the degree and direction of dominance of plus factors are 
similar for all loci,
iv. all plus factors are contributed by one parent and all minus factors contributed 
by the other parent,
v. no epistasis occurs among alleles at contributing loci and environmental, and
vi. genotypic variances are independent and combine additively to give total 
variability.
Lawrence and Frey (1976), suggested the use of the range of the F2 population in the 
estimation of effective factors when the parents did not represent the genotypic 
extremes for the segregating loci and when the variance of the F2 contained some 
non-additive components. Since cassava is a heterozygous crop, the Ft is genetically 
equivalent to an F2 of a homozygous crop (Magoon, 1967; Liu, 1998). Hence the 
formula used to estimate the minimum number of effective factors NE was;
where R was the range of the F1 segregates in a cross, cr2K is the genetic 
variance(Lawrence and Frey, 1976). The genetic variance cr,, was estimated from 
the linear function of the observed phenotypic variance and the environmental
66
variance, which was estimated from the homozygous population (Falconer and 
Mackay, 1997). In this study, the cassava parents are homogenous vegetatively 
propagated cuttings and no two plants of a single cloned genotype have the same 
phenotypic effect. The mean estimate of the variances of the two parents in a cross 
was, therefore, used as the environmental variance of homozygous population 
(Wricke and Weber, 1986; Falconer and Mackay, 1997).
Therefore, the genetic variance was
where Varg was the genetic variance; VarF1 was the F-i variance; VarPi was the 
variance of the better parent P-i, VarP2 was the variance of the worse parent P2 and n 
the number of observations in a parental mean. The number of factors affecting the 
CMD resistance trait contributed by each parent in a cross in each environment and 
across environments was estimated by using the procedure of Lawrence and Frey 
which is an extension of the Castle-Wright formula. Where little or no genetic 
variance was obtained, the number of effective factors was not estimated.
The number of favourable factors in the poorer parents were calculated as follows 
(Dixon et al., 1991)
where np was the number of plus (favourable) factors contributed by the poorer 
parent, X h = the mean score of the best Ft progeny of a cross, X  = mean score
of the better parent of a cross, X p - mean score of the poor parent of the cross 
X n,= mean score of the poorest F-i progeny of a cross. The number of favourable
67
(30)
factors in the better parent (nb) were obtained by subtracting np from NE, and 
because the number of factors estimated was the minimum, the number was 
approximated to the next larger interger.
3.3. Determination of DNA markers associated with resistance to CMD
Bulk segregant analysis (BSA) and linkage mapping were used to determine 
molecular markers linked to resistance to CMD in a resistant landrace using RAPD, 
SSR, and AFLP marker systems.
The BSA was carried out in three stages. In the initial screening, parental DNA and 
bulked DNA, were screened with 142 RAPD primers, 150 SSR, and 128 AFLP primer 
pairs. Primers that were polymorphic between the parents and the two bulks were 
selected for the second level screening. In the secondary screening, parents, bulks 
and the members of each bulk were screened with the selected primers. Primers that 
detected consistent differences between the resistant and susceptible DNA samples 
were then selected for the tertiary screening. The entire mapping population and the 
23 cassava clones used in the diallel and NCD II studies were screened.
3.3.1 Genetic material and experimental procedure
The mapping population consisted of 69 Fi progenies from a cross between the 
improved clone TMS I30555 (susceptible to CMD) and the resistant landrace TME 7 
(Oko-lyawo). The Ft progenies were developed and evaluated in the NCD II 
experiments described in section 3.1.1 and 3.1.3 above. In March 2000, cuttings of 
the progenies and parents were planted in nursery beds and watered twice a week to 
produce young leaves for DNA extraction. The 23 cassava genotypes used as 
parents and checks in the diallel and NCD II experiments (Table 3.1) were included 
to confirm the potential of a DNA marker association with CMD resistance as a 
molecular marker for resistance to CMD in the landraces.
68
3.3.2 Resistance screening and selection of plants for bulk segregant analysis
The plants were evaluated for symptom severity as described in section 3.1.3, at 6, 
12, 20, and at 50 WAP during the 1998/99 season and at 6 and 12 WAP during the 
1999 season in Ibadan. Disease severity scores of whole plants and shoot tips (from 
the shoot apex to the first fully expanded leaf) were assessed. At 12 WAP during the 
1999 growing season, the first 10 leaves of each plant were evaluated for symptom 
severity.
In March 2000, cuttings of the plants were made in a nursery bed to generate young 
leaves for DNA extraction. At 3 WAP, the shoot tips were excised for DNA extraction, 
and to enhance symptom severity. The plants were then assessed twice a week till 
the twelfth week for a final confirmation of resistance or susceptibility.
The mean disease severity scores based on shoot tip, whole plant, and the first 10 
leaves for the individual plants were determined from the data collected. Ten highly 
resistant (HR) Fi genotypes and ten highly susceptible (HS) genotypes of the 
mapping population were selected randomly to make up the two bulks for BSA.
3.3.3 DNA extraction
The DNA extraction kit, DNeasy™ from Qiagen was used for DNA extraction. The 
extraction was done according to the manufacturers’ instructions (Qiagen, 1999). 
Briefly, 100 mg of fresh leaves from the shoot apices was collected into 1 ml 
eppendorf tubes and ground to a fine powder in liquid nitrogen. The extraction buffer 
(400 pi buffer AP1) plus 4 |j| of Rnase (100 mg/ml) was added. The slurry was mixed 
with the pestle used to grind the sample; the sample was incubated at 65°C for 10 
minutes. The tubes were inverted three times during the incubation period to facilitate 
lyses of the cells.
69
After the incubation, 130 pi of the buffer AP2, which precipitates detergents, proteins 
and polysaccharides was added, mixed, and incubated on ice for 5 minutes, followed 
by centrifugation at 14,000 rpm for 5 minutes. The lysate was then applied to a 
QIAshredder spin column, and centrifuged for 2 minutes at 14,000 rpm to remove 
precipitates and cell debris. The flow-through was collected, and transferred to a 
fresh tube and 225 pi of buffer AP3 and 450 pi of ethanol was added and mixed 
gently to precipitate the DNA. The solution was then applied to the DNeasy mini spin 
column and centrifuged for 1 minute at 8,000 rpm to collect the DNA onto the matrix 
of the spin column. The column was then placed in a collection tube and washed 
twice with 500 pi buffer AW plus ethanol. The first wash was at 8,000 rpm for 1 
minute and the second wash at 14,000 rpm for 2 minutes.
The columns were allowed to dry at room temperature for about 30 minutes, then 
were placed in 1.5 ml eppendorf tubes. Pre-warmed eluting buffer (AE) was applied 
at 100 pi to each tube, incubated for 5 minutes at room temperature, and centrifuged 
for 1 minute at 8,000 rpm to elute the DNA. The elution step was repeated to give a 
total volume of 200 pi.
After extraction, the DNA quality was assessed on a 0.8% agarose gel in 0.5X tris 
borate EDTA (TBE) buffer (Appendix II). The DNA concentration was estimated with 
a Hoefer TKO 100 mini fluorometer and the purity checked by spectrophotometry.
3.3.4 DNA quantification
3.3.4.1. Agarose gel electrophoresis
Agarose solution of 0.8 % (w/v) concentration was prepared by weighing 1.6 g of gel- 
grade agarose (Sigma) into 200 ml of 0.5X TBE buffer. The emulsion was boiled in a 
microwave oven and allowed to cool to 50-55°C. For every 100 ml of agarose 
solution, 5 pi of ethidium bromide solution (10 mg/ml) was added. The solution was 
then poured into a horizontal gel electrophoresis tray (Owl E-D20) which was
70
mounted in a gel casting tray and fitted with two 50 teeth-combs (2 mm wide). After 
the gel had solidified, the tray was removed from the gel caster, and placed in its 
electrophoresis tank. The tank was filled with 900 to 1,000 ml of 0.5X TBE buffer, up 
to about 2 mm above the top of the gel, and then the combs were removed carefully 
from the moulded wells.
The extracted DNA samples were diluted (1:10) in tris EDTA (TE) buffer (Appendix 
III) and 10 pi aliquots mixed with 2.5 pi of loading dye (Appendix IV). Three 
concentrations of Arabidopsis DNA, were loaded with the samples as checks. The 
Arabidopsis DNA (100 ng/pl Gibco) was also diluted at 1:10 and 5 jjI, 10 pi and 15 pi 
aliquots mixed with 1/3 volume of loading dye and loaded alongside the samples.
The gel was run at a constant voltage (40V) till the marker dye was 3/4 down the 
length of the gel. The gel was then visualised under UV-light on a transilluminator, 
and documented with the aid of Grab-lt gel documentation computer software. The 
fluorescence intensities of the DNA samples were compared with those of the 
standards to give an indication of their concentration. The DNA samples were also 
checked for smearing. The presence of smearing indicated that the DNA had 
sheared during extraction; such a sample was discarded and fresh DNA of that 
sample extracted.
3.3.4.2 Estimation of DNA concentration and purity
The Hoefer TKO 100 mini fluorometer was used to estimate the concentration of the 
isolated DNA samples. It allows direct estimations of DNA concentrations in ng/pl by 
incorporating a dye, Hoechsts 33258 that binds DNA only. The fluorescence of 
Hoechsts 33258 in the presence of DNA depends on the AT content (Hoefer 
Scientific, 1992).
71
Prior to sample assay, the instrument was first switched on for 30 minutes to stabilise 
its readings. A standard DNA, 100 ng of Tomato or Arabidopsis DNA (Gibco BRL) 
was used to calibrate the instrument before samples were analysed. Each DNA 
sample (2 pi) was analysed in 2 ml of freshly prepared assay buffer (Appendix III).
Spectrophotometry readings at OO260 nm and OD280 nm were also taken to estimate the 
purity of the isolated DNA. The ratio of OD26o nm/OD2eo nm provides an estimate of the 
purity of the isolated DNA, which should be between 1.8 and 2.0 (Sambrook et al., 
1989).
3.3.5 Determination of CMD virus strain causing symptoms in the mapping 
population
The virus strain(s) causing the symptoms in the population was examined using the 
PCR method described by Zhou et al. (1997). DNA samples of the mapping 
population were tested with primers designed to identify the African cassava mosaic 
virus (ACMV), the East African cassava mosaic virus (EACMV) and both EACMV 
and the Uganda variant (UgV) of EACMV (Table 3.3).
Each 25|jl reaction was made up of 2 |jl DNA (1:40 dilution), 2.5 |jl 10 X PCR 
reaction buffer, 2.5 |jl MgCI2 (25 mM stock), 2.5 pi 5% Tween-20, 2.0 |jl dNTP's, 
dATP, dCTP, dGTP and dTTP (2.5 mM/|jl stock), 1 pi each of the forward and 
reverse primers (5.0 pM/pl), 1U thermus aquaticus (Taq) polymerase and sterile 
ultra-pure water.
The PCR reactions were performed in a 96-well thermal cycler PTC-200 DNA engine 
(MJ Research Inc. Watertown Massachusetts, USA.). The reaction cycles were as 
reported by Zhou et al. (1997). The first cycle consisted of 1 minute denaturing at 
94°C, 2 minutes primer annealing at 52°C, and 3 minutes extension at 72°C. This 
was followed by 35 cycles of 1 minute at 94°C, 1 minute at 52°C, and 1.33 minutes at
72
72°C. The final cycle consisted of 5 minutes at 72°C. PCR products were analysed 
on a 1.2% agarose gel in TBE as described in section 3.3.4.1, alongside a 1 kb DNA 
ladder (Gibco BRL, Gathersburg, Mel, USA).
73
Table 3.3 Nucleotide sequences of DNA primers used in polymerase chain reaction for
the detection of cassava mosaic begomoviruses strain in Fi mapping population
Primer Primer Type Forward and reverse Sequence
1 ACMV Specific ACMV-F1 5’-TTC AGT TAT CAG GGC TCG TAA-3'
ACMV-R1 5’-GAG TGC AAG TTG ACT CAT GA-3’
2 ACMV Specific ACMV-AL1/F 5’ GCG GAA TCC CTA ACA TTA TC 3'
ACMV-ARO/R 5’ GCT CGT ATG TAT CCT CTA AGG CCT G 3’
3 EACMV specific UV-AL1/F1 5’ TGT CTT CTG GGA CTT GTG TG 3'
EACMV-CP/R 5’ ACT CTA TGR GTA ATR CCY GA 3’
4 EACMV or UgV EACMV UV-AL1/F1 5’ TGT CTT CTG GGA CTT GTG TG- 3'
EACMV UV-AL1/R1 5' AAC CTA TCC CCG ATG CTC AT 3'
5 ug UV-AL1/F1 5’ TGT CTT CTG GGA CTT GTG TG 3’
ACMV-CP/R3 5’ TGC CTC CTG ATG ATT ATA TGT C 3'
74
3.3.6 Bulk segregant analysis
DNA samples of bulks, parents and individual FiS were standardised to 10 ng/|jl. For 
the AFLP analysis, bulks were prepared from successfully digested, ligated, and 
preamplified templates.
3.3.6.1 Random amplified fragment length polymorphism (RAPD) analysis
Equal aliquots (10 (jl) of each of the 10 DNA samples constituting a bulk group were 
combined in an eppendorf tube to make up the bulk DNA sample. A total of 142 
RAPD primers (Operon, Alamanda) were used to screen the parentals and bulks. 
Each 25 pi reaction contained 25ng DNA, 2.5 pi of the primer (2 pmol/pl), 2.5 pi of 
dNTP’s (2.5mM stock), 2.5 pi MgCI2 (25mM/pl stock), 2.5 pi 10x Taq polymerase 
buffer (Promega), 2.5 pi of 5% Tween-20, 2.5 pi DNA (10 ng/pl), and 1U Taq 
polymerase enzyme (Promega) in sterile ultra-pure water. PCR was performed in the 
Perkin Elmer, GeneAmp thermocycler. The cycling profile used for amplification 
consisted of a 3 minutes initial denaturation at 94°C followed by 45 cycles of 94°C for 
1 minute, 35°C for 1 minute, and 72°C for 2 minutes, a final extension for 7 minutes 
at 72°C, and then stored at 4°C until they were used. The PCR products were 
analysed on a 2% agarose gel in TBE as described in section 3.3.4.1 and 3.3.5.
3.3.6.2 Amplified fragment length polymorphism (AFLP) analysis
AFLP analysis was performed with three rare cutting restriction enzymes (6 base pair 
recognition sites), EcoR1, Apa1 and Pst1 and two frequent cutting (4 base pair 
recognition sites) restriction enzyme Mse1 and Taq1, to generate EcoRMMsel, 
Apa1ITaq1, and Pst1/Taq1 DNA fragments. The DNA fragments were ligated onto 
their respective adapters and used as templates for plus 1 and plus 3 amplification 
reactions. The sequences of the adapters, core primers sequences and the plus 1 or 
plus 3 extensions for the different restriction enzymes are given in Table 3.4.
7 5
Table 3.4. Nucleotide sequences of AFLP adapters, primers and primer extensions
Adapter and primer core sequences
Mse1 
Taq1 
Apa1 
EcoR 1 
Pstl
Adapters
5' GAC GAT GAG TCC TGA G 3'
5' TAC TCA GGA CTC AT 3'
5' GACGATGAGTCCTGAC 3 
5' CGG TCA GGA CTC AT 3’
5' TCGTAGACTGCGTACA GGCC -3'
5’ TGTACGCAGTCTAC 3'
5' CTC GTA GAC TGC GTA CC 3’
5' AAT TGG TAC GCA GTC 3'
5- CTCGTAGACTGCGTACATGCA-3 
5- TGTACGCAGTCTAC-3
Primers
+1 +3
Mse1 + E ‘5- GATGAGTCCTGAGTAA + E -3’ C CAA, CAC.CAG, CAT, CTA, CTC, CTG, CTT
Taq1 + E ‘5- CGATGAGTCCTGACCGA + E -3' A, C CAC, CAG, CAT, CTA, CTG, CTT, GAC, GTC
Apa1 + E ‘5- GACTGCGTACAGGCCC + E -3’ A, C ATT, CTA, CTG, GCA, TTG
EcoR1 + E ‘5- AGACTGCGTACCAATTC + E -3' A AAC, A AG, ACA, ACT, ACC, ACG, AGC, AGG
Pst1 + E '5- GACTGCGTACATGCAG + E -3’ A ACA, ACC, ACT, AGC
+1 or +3 nucleotide extension
76
For the EcoRMMsel analysis, the AFLP Analysis System I kit purchased from Life 
Technologies (Gibco BRL, Gathersburg, Mel, USA) was used. This kit is designed for 
use with plants having genomes ranging in size from 5. 3 x 10 8 to 6.3 x 109 bp and 
is suitable for cassava which has a genome size of 772 mega base-pairs (7.7 x 10 8) 
in the haploid genome (Awoleye et al., 1994).
For the Apa1!Taq1 and PstHTaql analysis, restriction enzymes, buffers and 
oligonucleotides were purchased separately from different companies (Roche and 
Phamarcia) and the reaction performed according to the methods described by Vos 
et al. (1995) and Wong et al. (1999).
Restriction of genomic DNA
EcoR1 !Mse1
The digestion mixes were set up in sterile 200 |j| PCR tubes. Each 25 pi was made 
up of 5 |jl of a 5X reaction buffer, 250 ng of each of DNA samples, 2.5 pi of £coR1 
and Mse1 enzyme mix (1 U/jjI) and AFLP-grade water. The tubes were flicked gently 
to mix the reaction mix, then spun down with a mini strip-tube centrifuge to collect the 
content. The digestion reaction was performed with the PTC-200 DNA Gene engine 
at 37°C for three hours. The restriction enzymes were then inactivated at 70°C for 15 
minutes, allowed to cool at room temperature, and then placed on ice prior to the 
ligation reaction.
Apa1ITaq1 and Pst1ITaq1
Each 50 |j| restriction reaction contained 500 ng DNA, 5 (jl 10X reaction buffer for the 
specific enzymes, 0.05 mM DTT, 10 ng BSA and 5 U each of enzyme. The Taq1 
digestion preceded the Apa1 or Pst1 digestion at 65°C for three hours, after which 
the Apa1 or Pst1 enzyme together with their incubation buffer was added and
77
digestion continued at 37°C overnight. All reactions were performed with the PTC- 
200 DNA Gene engine.
Ligation of adapters
EcoRMMsel
A 24 pi of adapter ligation solution plus 1 |jl T4 DNA ligase (1 U) was added to each 
digest giving a total reaction volume of 50pl. The tubes were mixed gently at room 
temperature, centrifuged briefly to collect the contents, and incubated at 20°C in the 
PCR machine for 2 hours. After the ligation, 10 pi aliquots of each sample was 
checked on a 1 % agarose gel to ensure complete digestion then stored at -20°C. 
The digestion and ligation were repeated for samples that showed evidence of 
incomplete digestion.
AoaHTaql and Pst1ITaa1
Prior to ligation, the adapters were prepared by adding equimolar amounts of their 
respective reverse and forward single strands in sterile ultra-pure water to the 
required working concentration. The double stranded adapters were verified on 1% 
agarose gel before use.
Each 10 pi ligation reaction mix was made up of 5 pmol of either Apa1 or Pst 1 
adapter and 50 pmol of the Taq1 adapter, 1 ul 10X one-phor-all buffer (Pharmacia), 
0.05 mM DTT, 10 ng BSA, 12 mM ATP, 1 U T4 DNA ligase and 2.5 U each of the 
restriction enzymes were added to each 50 pil restriction reaction. The reactions were 
then incubated at 37°C for three hours and the enzymes inactivated at 85°C for 30 
minutes. The samples were also checked on agarose gel as described for the Eco 
R1 IMsel reactions to ensure complete digestion and stored at -20°C.
Preamplification Reactions
FrnRHMsel
78
The preamplification reaction was performed with 5 (Jil of the undiluted restriction 
ligation, 40 pi of preamplification mix, 5 pi 10X PCR buffer plus MgCI2 (both supplied 
with the kit) and 1 U Taq DNA polymerase (Roche). The use of undiluted template for 
preamplification or passing the ligation mix through a column is known to improve the 
quality of selective amplification (Chavarriaga-Aguirre et al., 1999). The 51 pi 
reactions were covered with 10 pi of mineral oil and the preamplification was 
performed for 20 cycles at 94°C for 30 seconds, 56°C for 60 seconds, and 72°C for 
60 seconds in the PTC-200 Gene engine. At the end of the reaction, the block 
temperature was maintained at 4°C until the samples were removed from the PCR 
machine.
After the PCR reaction, 10 pi aliquots were run on 3% agarose gel to ensure 
amplification. The resistant and susceptible DNA bulks were then prepared by 
combining 10 pi each of the preamplified DNAs of the members of each bulk. The 
DNA templates for the selective amplification were then prepared as 1:20 dilutions of 
the preamplified DNA with TE buffer (supplied with the kit).
ApaHTaal and Pst1ITaq1
The Apa1ITaq1 fragments were preamplified with either, Apa1+CITaq1+A primers, 
Apa1+NTaq1+C primers, or Apa1+CITaq1+C primers. The Pst1ITaq1 fragments 
were amplified with Pst1+A/Taq1+C primers. Each 50 pi reaction contained 5 pi of 
the restriction/ligation reaction, 75 pmol each of the Pst1+1 or Apa1+1 and Taq1+1 
primers, 2 pi dNTP’s (5 mM stock, Promega), 5 pi 10X PCR buffer (Promega), 35 
mM MgCI2 (25 mM stock, Promega) and 1 U of Taq polymerase. The preamplification 
PCR reaction profile for the Apa1/Taq1 was 94°C for 2 minutes, followed by 35 
cycles of 94°C for 30 seconds, 50°C for 30 seconds, and 72°C for 60 seconds. At the 
end of the reaction, the block temperature was maintained at 4°C until the samples 
were removed from the PCR machine. The Pst1ITaq1 preamplification was
79
performed with the profile used for the EcoRMMsel reactions. The PCR reactions 
were then checked on agarose, diluted, and used to prepare bulk DNA as described 
for the EcoR'i/Msel reactions.
Selective amplification reactions
EcoRMMsel
Mse1 and £coR1 primers with a total of eight plus 3 nucleotide extensions each 
(Table 3.4) were used in all possible combinations (64) to screen the parents and 
bulked DNA. Each PCR reaction contained 2.5 pi of the 1:20 dilution, 2.0 pi I of 10X 
PCR buffer (plus 15 mM MgCI2l), 0.6 pi MgCI2 (25 mM stock), 0.5 pi EcoR1 primer 
(10 ng/pl), 4.5 pi of Mse 1 primer mix (6.7 ng/ul with 0.9 mM dNTP’s) and 0.5 U Tag 
polymerase (Roche).
The selective amplification was carried out for one cycle at 94°C for 30 seconds, 
65°C for 30 seconds, and 72°C for 60 seconds. This was followed by 12 cycles 
where the annealing temperature was lowered by 0.7°C for each cycle, giving a touch 
down phase of 13 cycles, then continued with 23 cycles at 94°C for 30 seconds, 56°C 
for 30 seconds, and 72°C for 60 seconds. At the end of the reaction, the block 
temperature was maintained at 4°C until the samples were removed from the PCR 
machine.
Aoa1/Taq1 and Pst1/Taa1
The Apa1+CITaq1+A templates of the parents and bulks were screened with all 
possible combinations of the Apa1+3 and Taq1+3 primers (Table 3.4). The 
Apa1+AITaq1+C templates were screened with six of the primer combinations 
involving Apa1+ATT and Taq1+CAC, Taq1+CAG, Taq1+CAJ, Taq1+CTA, Taq1+ 
CTG or Taq1+ CTT. The Apa1+CITaq1+C templates were screened with 12 of the
80
primer combinations involving Apa1 + CTA or CTG and Taq1 + CAC, CAG, CAT, 
CTA, CTG or CTT.
The Psf1+A/7ag1+C templates were screened with 24 of the primer combinations 
involving Pst1+ ACA or Pst1+ ACC and Taq1+CAC, Taq1+CAG, Taq1+CAT, 
Taq1+CTA, Taq1+CTG or Taq1+CTT. Each 20ul PCR reaction mix contained 2.5 pi 
of the 1:20 diluted preamplification reaction, 2.0 pi 10X PCR buffer, 2.0 pi MgCI2 (25 
mM/pl), 1 pi Apa1 or Pst1 plus three primers (5 ng/pl stock), 1 pi Taq1 primers (30 
ng/pl stock), 0.8 pi dNTP’s (5 mM stock) and 0.5 U Taq polymerase (Promega).
The selective amplification for the Apa1ITaq1 templates was carried out for one cycle 
at 94°C for 60 seconds, 65°C for 30 seconds, and 72°C for 60 seconds. This was 
followed by 12 cycles where the annealing temperature was lowered by 0.7°C for 
each cycle, giving a touch down phase of 13 cycles. This was followed by 23 cycles 
at 94°C for 30 seconds, 56°C for 30 seconds, and 72°C for 60 seconds. At the end of 
the reaction, the block temperature was maintained at 4°C until the samples were 
removed from the PCR machine. The profile for the PstHTaql was the same as that 
used for the EcoR1/Mse1 reactions. The PCR products were analysed immediately 
or stored at -20°C for a few days before analysis.
3.3.6.3 Simple Sequence Repeats (SSR) analysis
A total of 186 SSR previously reported (Chavarriaga-Aguirre et al., 1998; Mba et al, 
2001) were used to screen the parents and two bulks. Each 25 pi reaction contained
2.5 pi DNA (10 ng/ul), 2.5 pi each of the forward and reverse primers (2.0 uM each), 
1 pi dNTPs (5 mM stock), 2.5 pi 10X PCR buffer, 1 or 1.5 pi MgCI2 (25 mM) and 
0.75-0.8 units of Taq polymerase. The PCR profile involved an initial denaturation for 
5 minutes at 94°C. The first cycle then had a denaturation at 94°C for 30 seconds, 
annealing at 65°C for 1 minute, and extension at 72°C for 1 minute. This was 
followed by 10 cycles, which were similar to the first cycle, but the annealing
81
temperature was lowered by 0.7°C for each cycle. Thereafter, the annealing 
temperature was maintained at 55°C for 1 minute, for 25 cycles. The final extension 
was 72°C for 7 minutes. At the end of the reaction, the block temperature was 
maintained at 4°C until the samples were removed from the PCR machine.
3.3.6.4 Sequencing gel electrophoresis
AFLP selective amplification products and SSR amplification products were analysed 
on 6% polyacrylamide (Appendix IV) sequencing gels, using either an Owl S2S 
sequencing apparatus with gel size 35 X 45 cm or a Gibco Life Technologies Model 
S2 with gel size 31.0 X 38.5 cm.
Prior to sequencing gel electrophoresis, a pair of clean glass plates (a long and short 
plate) was washed with soap, followed by several rinses with tap water, then a 10-30 
minute rinse in deionised water with agitation. When the plates were dry, they were 
cleaned with absolute ethanol. One plate was treated with 400 ul of a siliconizing 
agent (Appendix IV) for 5 minutes then the excess was wiped off gently with about 2 
ml ethanol. The other plate was also swabbed with alcohol then treated with 1 ml of a 
gel binding solution (Appendix IV). The plates were then assembled on a gel caster 
with a pair of 0.4 mm spacers placed between them and clamped along their vertical 
sides. For the 31.0 X 38.5 cm gel, about 70 ml of the 6% polyacrylamide gel 
(Appendix IV) was poured between the two plates and for the 35 X 45 cm gel about 
90 ml was used. When the gel had filled the plates, the flat surface of a 0.4 mm 
shark-tooth comb was inserted, clamped, and left to polymerise for at least 1 hour.
When the gel had polymerised, the clamps and comb were removed; the gel cassette 
was washed with deionised water and mounted on the appropriate sequencing gel 
apparatus and the upper and lower trays of the tank were filled with about 500 ml of 
1X TBE. With a 50 ml syringe, the moulded well was washed with the running buffer 
to remove gel pieces and excess urea. The tank was closed and the gel prewarmed
82
at 60 W for at least 30 minutes till the temperature of the gel was 50°C. After pre­
warming, the wells were flushed again with buffer to get rid of urea, and then the 
combs were fixed in the wells. Thereafter, the samples were loaded (within 15 
minutes).
Prior to loading, 1/3 volume of formamide denaturation dye (Appendix IV) was added 
to each PCR product, denatured for 3 minutes at 95°C, and immediately placed on 
ice. About 8 |jl of denatured sample was applied to the combs. A 50 bp marker 
(Promega) with 16 DNA fragments ranging from 50 to 800 bp in exactly 50 bp 
increments was loaded alongside the samples. For the 31.0 X 38.5 cm gel, 48 
samples including the marker were loaded per gel and 72 for the 35 X 45 cm gel. The 
electrophoresis was continued at 60 W with the gel temperature at 50°C for 1.5-2 
hours for the 31.0 X 38.5 cm gel, and 2.5-3 hours for the 35 X 45 cm gel.
3.3.6.5 Silver staining
The AFLP bands were visualised using the commercially available silver staining kit 
(Promega part No. Q4132). The silver staining procedure involves three stages. A 
fixing stage removes electrophoresis buffer and urea from the gel and prevents 
diffusion of small extension products. This is followed by staining the gel in a solution 
of AgN03 (0.1 %). It is then developed in a staining solution made of Na2C03 (3 %) in 
the presence of formaldehyde and sodium thiosulphate which reduces the silver into 
metallic silver (Promega, 2000). The procedure was carried out according to the 
manufacturer’s instructions. Reagents used for silver staining are listed in Appendix I.
After electrophoresis, the combs and spacers were removed and the two plates 
separated. The gel was placed (gel side up) in a plastic tray and enough fixer was 
added to completely cover the gel, and agitated gently for 20 minutes. This was 
followed with three washes in deionised water for 3 minutes each. The plate was 
then transferred to another tray containing enough stainer to completely cover the gel
83
and agitated for 30 minutes on a shaker. The gel was then washed briefly (5 to 10 
sec) in deionised water to wash off excess silver, and transferred to a tray containing 
cold developer. The gel was then developed until the bands were visible, and 
transferred back into the fixer for 2 minutes to stop the reaction. It was then washed 
twice in deionised water, placed in an upright position, and allowed to dry.
3.3.6.6 Gel analysis, marker scoring and nomenclature
Agarose gels were documented under UV light with the Grabit gel documentation 
software and then scored with the Gelworks advanced documentation system. 
Polyacrylamide gels were also documented directly from the glass plates with the 
Grabit gel documentation software or by scanning. The sequencing gels were scored 
directly on the glass plate with the aid of a light box.
For the primary screening of the parents and bulks, the total number of bands and 
the number of polymorphic bands, i.e., bands clearly present in one parent or bulk 
and absent in the other, were recorded. In genotyping the mapping population, only 
the polymorphic bands were scored. Each lane containing DNA from a specific F-, 
plant was scored for the presence or absence of polymorphic bands based on its 
similarity with the parents. When the polymorphic band was present in TMS I30555 
(P-i) and absent in TME7 (P2), the marker was scored as ‘1’ for F-, plants that 
possessed the marker band, while those F-, plants that lacked the marker band were 
scored as ‘2’. Conversely, when the polymorphic band was absent in P-, but present 
in P2, the marker was scored as T  for Ft plants that lacked the band; and '2’ for Ft 
individuals that possessed the band. Faint bands that could not be scored with 
confidence were treated as missing data and were assigned scores of 'O'. When the 
NCDII and diallel parental and check clones were scored, polymorphic markers were 
scored as “1” for present or ”0” for absent.
84
m
t .H i
The marker sizes were estimated by extrapolating their sizes from the migration 
distance of the 50 bp ladder for the AFLP and SSR gels and from the 1 kb marker for 
the RAPD gels. Markers were named according to the primer from which the marker 
was obtained, and its relative molecular weight. For example, AFLP marker E- 
ACC/M-CTC-250 indicated that the marker was amplified with the primer combination 
of £coR1+ACC and Mse1+CTC and its molecular weight relative to the 50 bp marker 
was 250 bp. SSR markers were identified based on their assigned lab number and 
their expected molecular weight. When more than one allele was present, the alleles 
were identified with alphabets.
3.3.7 DNA markers associated with CMD resistance in Fi population
3.3.7.1 Genetic linkage mapping
After the initial screening for the BSA, primers that were polymorphic between the 
bulks and parents were used to genotype the entire mapping population. The 
EcoRMMsel +3 primers (E-ACC/M-CTC, E-AAC/M-CAT, E-ACA/M-CAC, E-AAG/M- 
M-CTC, E-ACA/M-CAT, E-AAG/M-CTG, E-ACT/M-CAG, E-ACC/M-CAT, and E- 
ACC/M-CTC) were used to genotype the parents and Fi population. Three PstHTaql 
primers P-ACC/T-CTA, P-ACT/T-CTT, and P-ACC/T-CAG and 15 SSR (Table 3.5.) 
primers were also used to genotype the parents and Fi population, to generate data 
for linkage analysis. PCR reactions, gel electrophoresis, developing and documented 
were done as described in sections 3.3.6.4, 3.3.6.5 and 3.3.6.6 above.
For the SSR analysis, PCR products of the mapping population were multiplexed in 
sets of three for each gel according to their base pair (bp) expected product sizes, 
with the products with a smallest bp loaded first and that with the largest loaded 
last.The CMD severity classes of the progenies were treated as a putative 
phenotypic locus designated CMDRL and included in the linkage analysis with the 
DNA markers such that the relationship between resistance to CMD and the markers
8 5
could be identified on the map. Plants that were classified as resistant (CMD scores 
of 1 and 2) were treated as identical to the resistant parent and were assigned a 
score of 2 while those that were classified as susceptible (CMD scores 3 to 5) were 
assigned a score of 1.
The markers were subjected to chi-square analysis to test their conformity to a 1:1 
inheritance ratio to identify single-dose restriction fragments (SDRFs). SDRFs are 
DNA markers that are present in one parent and absent in the other and segregate in 
a 1:1 ratio (absence: presence) in the progeny. They represent the segregation 
equivalent of an allele at a heterozygous locus in a diploid or an allopolyploid 
genome or a simplex allele in an autopolyploid (Wu et al., 1992). SDRFs are suitable 
for linkage analysis in an F-i population when there is the presence of a number of 
unique segregating polymorphisms (heterozygosity) and normal meiosis in either or 
both parents. It results in two separate linkage maps based on male and female 
(Fregene et al., 1997; Williams, 1998). The remaining markers, which did not 
conform to the 1:1 ratio, were then tested for their conformity to 3:1 ratio, which is 
expected for the segregation of double dose markers on homoelogous chromosomes 
of an alio or autopolyploid (Wu et al., 1992). Double dose markers represent the 
duplex (double simplex condition) at a heterozygous locus.
Different strategies for linkage analysis were tested. These were generating separate 
female and male maps which is valid for heterozyous species (Fregene et al., 1997; 
Williams, 1998); single combined map, inclusion of inverted scores of the female and 
male mapping data before linkage analysis (Fregene et al., 1997), and tested for both 
separate and combined maps.
The data were analysed with the MAPMAKER version 3.0 computer programme 
(Lander et al., 1987), using the markers generated in the study. Three CMD 
responses; mean shoot tip severity score (MST), mean whole plant severity score
86
(MWP) and mean leaf severity score (MLS) were included in the data set as three 
quantitative traits. The Kosambi mapping function (Kosambi, 1944) was employed for 
converting recombination fractions to map distances in centiMorgans (cM). A two- 
point analysis was first performed to assign loci to groups using minimum lod scores 
of 3.0 and maximum recombination fraction frequency of 0.33. This function inferred 
linkage between markers. The multipoint likelihood function was used to compare loci 
order in a linkage group. Initially a subset of markers was compared to determine the 
best order. Then the remaining markers in the group were tried in every interval using 
the 'multipoint/try' function. Locus order was then produced using the MAP function of 
MAPMAKER. Unlinked markers were further assessed by 'two-point/pairwise' 
function with reference to the already ordered linkage groups. The linkage groups 
were then assigned onto chromosomes and a framework map generated for further 
analysis.
8 7
Table 3.5 Description of cassava SSR loci and their primer pairs used in genetic linkage analysis
Lab No SSR loci Type of repeat Forward primer Reverse primer Size (bp) MgCI2
SSR6 SSRY6 (CA) 7 (N) 51 (CA) 17 (N) 47 (CA) 15 TTTGTTGCGTTT AGAAAG GT GA AACAAATCATTACGATCCATTTGA 298 1.5
SSR7 SSRY7 (CT) 26 TGCCT AAGGAAAATT CATT CAT TGCT AAGCT GGT CAT GCACT 250 1.5
SSR30 SSRY28 (CT) 26 (AT) 3 AC( AT) 2 TTGACATGAGTGATAI I I ICTTGAG GCTGCGTGCAAAACTAAAAT 180 1.5
SSR41 SSRY40 (GA) 16 TGCATCATGGTCCACTCACT CATTCI I I ICGGCATTCCAT 231 1.5
SSR50 SSRY49 (GA) 25 T GAAAAT CTC ACT G GC ATTATTT TGCAACCATAGTGCCAAGC 300 1.5
SSR52 SSRY51 (CT) 11 CG(CT) 11 (CA) 18 AG GTTG G AT GCTT G AAG GAA GGAT GCAGGAGTGCTCAACT 298 1.5
SSR103 SSRY91 (GA) 16 GTCTGCATGGCTCGATGAT TGCCTGCTTCATATGTT'I I TG 300 1
SSR108 SSRY95 (CT) 19 (CA) 16 CC (CA) 2 CC( CA) 3 CAT G ATTT G G ATTTT GG AAT GA CAAAAGAAGCAACCTT CAGCA 282 1
SSR115 SSRY102 (GT) 11 TTGGCTGCTTTCACTAATGC TT GAACACGTT GAACAACCA 179 1.5
SSR119 SSRY106 (CT) 24 GGAAACTGCTTGCACAAAGA CAGCAAGACCAT CACCAGTTT 270 1.5
SSR124 SSRY110 (GT) 12 TT GAGTGGT GAAT GCGAAAG AGTGCCACCTTGAAAGAGCA 247 1.5
SSR132 SSRY113 (GA) 19 TTTGCTGACCTGCCACAATA T CAACAATT G GACTAAG CAG C 187 1.5
SSR157 SSRY135 (CT) 16 CCAGAAACT GAAATGCAT CG AACATGTGCGACAGTGATTG 253 1.5
SSR70 SSRY170 (TA) 5 (N) 71 (CT) 24 T CT CGATTTGGTTT GGTT CA TCATCCTTGTTGCAGCGTTA 299 1.5
SSR105 SSRY179 (GA) 28 CAGGCTCAGGTGAAGTAAAGG GCGAAAGTAAGTCT ACAACTTTT C 
TAA
226 1.5
(Adapted from Mba etal., 2001)
88
3-3.7.3 Quantitative trait loci (QTL) analysis
Simple phenotypic correlation and regression analysis were used to determine the 
association between the markers and the CMD responses. The CMD responses, 
(mean shoot tip, mean whole plant, and mean leaf severity scores at 12 WAP) were 
the dependent variables while the marker data were the independent variables for 
the regression analysis. Associations between markers and the CMD response traits 
were declared significant on a per marker basis at a significance threshold of p = 
0.005 (Lander and Botstein, 1986). This stringent threshold was adopted to avoid 
type 1 error (Dudley, 1993).
3.3.8 DNA markers associated with CMD resistance in various sources of 
resistance
Following the segregation and QTL analysis, markers associated with resistance to 
CMD were tested on the 23 cassava genotypes used as parents and checks in the 
NCD II and Diallel experiments described in section 3.1 above.
89
Chapter Four 
Results
4.1 Analysis of variance and combining ability analyses
4.1.1. 3X18 North Carolina Design II mating scheme
The analysis of variance for CMD symptom severity in the NCD II mating scheme is 
presented in Table 4.1. The results showed significant variation (p<0.01) among 
environments, replicates nested in environments and genotypes in each of the 
individual environments and across environments. Significant variation among the 
genotypes was due to the variation among checks and among test genotypes. The 
orthogonal contrasts between the checks and test genotypes were significant 
(p<0.01) in the individual environments but not across environments.
The significant variation among the test genotypes was due to significant variation 
(p<0.01) among the parents and crosses in Ibadan 1998 and 1999 and across 
environments. In the Mokwa 1998 environment, significant variation among the test 
genotypes was due to the parents only.
Variation among the 21 parents was due to the male parents, which varied 
significantly (p<0.01) in each environment and across environments, as well as to the 
female parents, which was significant only in the individual environments. The 
orthogonal contrast female versus male parents was significant (p<0.05) in Mokwa 
1998 and across environments.
90
Table 4.1 Analysis of variance for CMD severity at 12 weeks after planting (WAP)
among all genotypes in 3X18 NCD II mating scheme, evaluated in three environments
Source of Variation
Across
Environment
df MS df (t)
Individual
Mokwa
1998
MS
Environments 
Ibadan Ibadan 
1998 1999 
MS MS
Environment (E) 2 59,17“
Replicates within E 3 1.40** 1 1.18** 2.98** 0.05’ *
Genotypes in population (G) 78 1.39** 78(77) 0.36** 0.49** 1.06**
Checks (Ck) 3 7.72** 3(2) 2.46** 2.71** 3.48**
Test Genotypes (TG) 74 1.12** 74 0.30** 0.38** 0.96**
Ck versus TG 1 1.32 1 2.57** 1.97** 1.56**
Parent (P) 20 3.17** 20 0.91** 0.89** 1.90**
Female (F) 2 164 2 0.55** 0.38* 1.24**
Male (M) 17 3.51** 17 0.99** 0.99** 2.08**
F versus M 1 0.46* 1 0.23* 0.07 0.19
Susceptible (S) 5 1.08 5 0.36** 0.68** 0.51**
Resistant (R) 14 0.29 14 0.11** 0.16* 0.63**
R versus S 1 53.98** 1 15.15** 12.55** 26.47**
Crosses (C) 53 0.31** 53 0.04 0.19** 0.37**
F (GCA) 2 1.40 2 0.18** 2.28** 0.42**
M (GCA) 17 0.57* 17 0.06* 0.14 0.91**
F x M (SCA) 34 0.11* 34 0.02 0.09 0.13**
P vs C 1 3.10 1 1.77** 0.54* 13.22**
GxE 155 0.28**
Ck xE 5 0.07
TG x E 148 0.26**
Ck vs TG x E 2 2.71**
P xE 40 0.26**
F xE 4 0.26**
M xE 34 0.28**
F vs M x E 2 0.14
S x E 10 0.36**
R xE 28 0.36**
R vs S x E 2 1.01**
C xE 106 0.15**
F (GCA) x E 4 0.54*
M (GCA) x E 34 0.27*
F x M (SCA) x E 68 0.07
P vs CxE 2 6.20**
Error (genotypes) 233 0.07 78(77) 0.04 0.09 0.07
Error (crosses) 159 0.06 53 0.03 0.08 0.08
GCA 0.76 1.00 0.94 0.84
GCA + SC/1
(t) df of genotypes and checks in Mokwa 1998 was 77 where check, 91/02324 was missing 
* Significantly different from zero at the 0.05 probability level 
** Significantly different from zero at the 0,01 probability level
91
Further partitioning of the parents into resistant and susceptible parents revealed 
significant differences among parents for each group in the individual environments 
but not across environments. The orthogonal contrast between resistant and 
susceptible parents, however, was significant (p<0.01) in all three environments and 
across environments
The reaction of the Ft crosses to CMD symptom severity was significant (p<0.01) in 
Ibadan 1998 and 1999, and across environments. This was attributed to the GCA 
effects of the females in the Ibadan 1998 and 1999 environments and the GCA 
effects of the males in Ibadan 1999 and across environments. The SCA effect 
(female by male interaction) was also significant across environments and in Ibadan 
1999. Although there was no significant variation among the crosses in Mokwa 1998, 
the GCA effect of the females and males was significant.
The orthogonal contrast of the parent versus crosses, which is a test of average 
heterosis, was significant in the individual environments but not across environments.
The significant GXE interaction (p<0.01) was due to all the component interactions 
except for the check by environment, the parent contrast, female versus male by 
environment, and the SCA effect (female by male) by environment interactions.
The relative importance of GCA and SCA estimated by the genetic ratios,
GCA-------------were 0.76, 1.0, 0.84 and 0.96 across environments, in Mokwa 1998,
GCA + SCA
Ibadan 1998 and Ibadan 1999 environments respectively. Heritability based on 
progeny means across environments was 0.516 with lower and upper 90% 
confidence interval of 0.268 and 0.668.
92
4.1.2. 4X4 Diallel mating scheme
The analysis of variance for the 4X4 diallel involving the resistant genetic stock clone 
58308 and three improved cassava clones (130001, I30555 and I30572) also used as 
females in the 3X18 NCD II mating is presented in Table 3.2.
There was significant variation (p<0.01) among environments but replicate within 
environment was significant only in Mokwa 1998. The variation due to test genotypes 
was also significant (p<0.01) in the individual environments and across 
environments. This was attributed to the significant variation due to checks and test 
genotypes. The contrast check versus test genotypes was significant in the Mokwa 
1998 environment only.
The results further showed that significant variation (p<0.01) among the test 
genotypes in all the three environments and across environments was due to 
variation among parents and crosses. The test for average heterosis, parent versus 
cross contrast, was significant in all of the three environments but not across 
environments.
The Griffing’s analysis of variance (based on method 1 for a complete diallel, and 
model 1 for fixed genotypes) for the crosses within and across environments 
revealed that GCA effects of the parents were significant in the individual 
environments but not across environments. SCA effects also contributed significantly 
to the variation among crosses in the Ibadan 1999 and Mokwa 1998 environments. 
Reciprocal effects were significant only in the Ibadan 1998 environment and this was 
due to significant maternal effect.
93
Table 4.2 Analysis of variance for CMD severity at 12 weeks after planting (WAP) among
all genotypes in 4X4 diallel mating scheme, evaluated in three environments
Source of Variation ' Across Individual Environments
Environment Mokwa Ibadan Ibadan
1998 1998 1999
df MS df MS MS MS
Environment (E) 2 9.00**
Replicates within E 3 0.14 1 0.25** 0.08 0.08
Genotypes in population (G) 23 1.47** 23(21)$ 0.46** 0.67** Q.78**
Checks(Chk) 3 6.74** 31 2.72** 3.26** 1.95**
Test Genotypes (TG) 19 0.65** 19(18) $ 0.14** 0.27** 0.62**
Parent (P) 3 1.08* 3 (2 )* 0.11** 0.51** 0.81*
Cross (C) 15 0.37* 15 0.15** 0.17** 0.37*
GCA 3 0.79 3 0.47** 0.39** 0.51*
SCA 6 0.40 6 0.1** 0.04 0.56*
Reciprocal 6 0.13 6 0.05 0.19* 0.12
Maternal 3 0.11 3 0.06* 0.36** 0.06
Non-matemal 3 0.15 3 0.03 0.02 0.17
P v s C 1 2.36 1 0.09* 1.33** 3.73**
Chk vs TG 1 0.36 1 1.69** 0.14 0.56
GxE 43 0.21*
Chk xE 5 0.17
TG x E 36 0.19**
P x E 4 0.11
C x E
GCA x E
SCA x E
R x E
M x E
N x E
30
3
6
6
3
3
0.16**
0.47**
0.23
0.21
0.35*
0.06
P vs C x E 1 1.16**
Chk vs TG x E 1 1.09**
Error (genotypes) 66 0.11 23(21 )* 0.02 0.08 0.23
Error (cross) 45 0.08 15 0.02 0.05 0.16
2 GCA 0.19 0.39 0.76 0.10
2GCA + SCA
( )  df of genotypes, test genotypes and parents in Ibadan 1998 are 21, 18 and 3 respectively where 58308 
was missing
t  df of genotypes, test genotypes checks and parents in Mokwa 1998 are 20, 17 and 2 respectively where 
58308 and 91/02324 were missing 
* Significantly different from zero at the 0.05 probability level 
** Significantly different from zero at the 0.01 probability level
94
Despite the lack of significance of the reciprocal effect in Mokwa 1998, significant 
maternal effects were detected. The partitioning of the reciprocal effects into its 
components revealed significant maternal effect in the Mokwa 1998 and Ibadan 1998 
environments.
The significant genotype by environment interaction (p<0.05) was due to the 
interactions between environment and test genotypes, crosses, GCA effect and 
maternal effects, parent versus cross and the check versus test genotypes.
The relative importance of GCA and SCA estimated by the genetic ratios,
2Xj (2A---------------were 0.19, 0.39, 0.76 and 0.10 for the combined environment analysis,
2 GCA + SCA
the Mokwa 1998, the Ibadan 1998 and the Ibadan 1999 environments respectively. 
Heritability based on progeny means was 0.568 with lower and upper 90% 
confidence interval of 0.028 and 0.786.
4.1.3 Combining ability effects of 3X18 NCD II mating scheme
The GCA effects and means of the parental clones in the three environments and 
across environments are presented in Table 4.3. Mean disease severity scores of the 
clones varied across and within each environment.
In this study, negative GCA effects were desirable for resistance. Significant and 
negative values of the clones indicated contribution towards resistance while 
significant and positive values indicated a contribution towards susceptibility. 
Similarly, negative SCA values of a cross indicated that the cross was more resistant 
than average and significant and positive SCA effect indicated a cross was more 
susceptible than average.
95
Table 4.3 Mean* CMD severity scores and GCA effects for CMD symptom severity of
parents in the 3X18 NCD II mating scheme, evaluated in three environments
Parent
Across 
environments 
Mean+ GCA
Mokwa 1998 
Mean GCA
Ibadan 1998 
Mean GCA
Ibadan
Mean
1999
GCA
Female
130001 1.20 0.13 1.00 0.08 1.15 0.29* 1.45 0.02
I30555 2.23 -0.04 1.95 -0.02 1.85 -0.11 2.88 0.02
I30572 1.55 -0.09 1.10 -0.06 1.94 -0.17* 1.60 -0.04
Mean 1.66 1.35 1.65 1.98
SE 0.12 0.06 0.25 0.02 0.22 0.04 0.18 0.04
LSD 0.258 0.165 0.165 0.083
Male
14(2)1425 1.96 -0.06 1.20 -0.04 1.87 0.16 2.80 -0.31*
160142 1.63 -0.20 1.00 -0.19* 1.30 -0.06 2.60 -0.34*
I90257 1.48 -0.26* 1.30 -0.14 1.54 -0.01 1.60 -0.63*
TME1 1.48 -0.09 1.10 -0.04 1.58 -0.07 1.75 -0.17
TME2 2.68 0.55* 2.85 0.23* 2.13 0.44* 3.06 0.98*
TME4 1.18 -0.02 1.00 0.02 1.21 -0.08 1.34 -0.01
TME5 1.20 0.06 1.10 0.04 1.20 0.14 1.30 0.01
TME6 1.37 -0.07 1.25 -0.04 1.48 -0.19 1.38 0.01
TME7 1.12 -0.03 1.10 0.05 1.11 -0.05 1.15 -0.07
TME8 1.36 -0.12 1.30 0.00 1.08 -0.06 1.70 -0.29*
TME9 1.15 -0.05 1.05 0.14 1.25 -0.03 1.15 -0.27*
TME10 2.76 0.05 1.93 -0.01 2.78 0.09 3.55 0.07
TME11 1.44 -0.09 1.29 0.12 1.73 -0.22 1.30 -0.18
TME12 1.39 -0.09 1.30 0.03 1.43 -0.12 1.45 -0.20
TME14 1.37 0.06 1.10 0.03 1.30 0.09 1.70 0.06
TME31 2.77 0.20 3.05 -0.03 2.17 -0.04 3.10 0.67*
TME41 3.15 0.01 2.55 -0.10 3.05 -0.11 3.85 0.24*
TME117 3.46 0.17 2.65 -0.05 3.49 0.09 4.25 0.48*
Mean 1.83 1.56 1.76 2.17
SE 0.12 0.06 0.25 0.02 0.22 0.04 0.18 0.04
LSD 0.238 0.235 0.235 0.144
* Significantly different from zero at the 0.05 probability level 
** Significantly different from zero at the 0.01 probability level 
*mean disease score < 2  = resistant; mean disease scores > 2=susceptible.
96
Negative GCA effects were estimated for the resistant clone TMS 130572 in all 
environments but its effect was significant only in the Ibadan 1998 environment. 
Although the resistant clone TMS 130001 had positive GCA in all environments and 
across environments, its effect was significant only in Ibadan 1998. All the improved 
clones used as males TMS 14(2)1425, TMS 160142 and TMS I90257 had negative 
and significant GCA effects in Ibadan 1999. Negative and significant GCA effects 
were also estimated for clones TMS 160142 in the Mokwa 1998 environment and for 
TMS I90257 across environments. Negative and significant GCA effects were also 
detected for the resistant landraces TME8 and TME9 in the Ibadan 1999 
environment.
The susceptible landrace TME2 had significant and positive GCA in all three 
environments and across environments, while the susceptible landraces TME31, 
TME41 and TME117 had significant and positive GCA effects only in the Ibadan 
1999 environment. The susceptible improved clone TMS I30555 used as female, did 
not contribute significantly to susceptibility in any environment.
The correlation analysis between parental means and GCA revealed significant and 
positive linear relationships in the Ibadan 1999 environment (r=0.551, p<0.05) and 
across environments (r=0.507, p<0.01).
The mean disease severity scores and specific combining ability effects of the 
crosses for environments with significant variation for the effect are presented in 
Table 4.4. Negative SCA effects were also desirable for resistance. The most 
resistant cross over environments was TMS I30572 X TMS I90257 which had the 
lowest mean of 1.48 and largest negative SCA effect of -0.43 while the TMS I30555 
X TME2 with a mean of 2.57 and largest positive SCA of 0.63 was the most 
susceptible cross across environments.
97
Table 4.4 Mean CMD severity scores and SCA effects for CMD symptom severity of
crosses in 3X18 design II crosses in environment with significant variation for SCA
Across Environments Ibadan 1999 
Cross Mean SCA Mean SCA
130001 X 14(2)1425 2.09 -0.03 2.34 -0.17
130001 X 160142 1.84 -0.28* 2.12 -0.36*
130001 X I90257 1.96 -0.16 2.48 0.30
130001 X TME1 1.95 -0.17 2.25 -0.40*
130001 X TME10 2.2 0.08 2.99 0.10
130001 X TME11 1.92 -0.20 2.45 -0.19
130001 X TME117 2.12 0.00 2.98 -0.31
130001 X TME12 1.91 -0.21 2.64 0.02
130001 X TME14 2.28 0.16 3.23 0.35*
130001 X TME2 2.7 0.58* 3.81 0.02
130001 X TME31 2.36 0.24* 3.72 0.23
130001 X TME4 2.2 0.08 2.89 0.08
130001 X TME41 2.21 0.09 3.25 0.19
130001 X TME5 2.09 -0.03 2.66 -0.16
130001 X TME6 2.09 -0.03 2.8 -0.03
130001 X TME7 2.2 0.08 2.92 0.19
130001 X TME8 2 -0.12 2.47 -0.06
130001 X TME9 2.04 -0.08 2.75 0.20
I30555 X 14(2)1425 2 0.05 2.78 0.27
130555 X 160142 1.74 -0.21 2.65 0.17
I30555 X I90257 1.74 -0.21 2.04 -0.15
I30555 X TME1 1.8 -0.09 2.56 -0.09
I30555 X TME10 2.14 0.13 2.98 0.09
I30555 X TME11 2.04 0.09 3.03 0.39*
I30555 X TME117 2.17 0.21 3.4 0.10
I30555 X TME12 1.69 -0.24* 2.49 -0.13
I30555 X TME14 1.97 0.02 2.78 -0.10
I30555 X TME2 2.57 0.62* 3.89 0.09
I30555 X TME31 2.06 0.11 3.36 -0.14
I30555 X TME4 1.78 -0.17 2.51 -0.30
I30555 X TME41 1.83 -0.13 2.85 -0.22
I30555 X TME5 2.23 0.28* 3.14 0.30
I30555 X TME6 1.93 -0.02 2.96 0.13
I30555 X TME7 1.84 -0.13 2.58 -0.20
I30555 X TME8 1.75 -0.20 2.44 -0.09
130555X TME9 1.83 -0.13 2.4 -0.15
SE
LSD
0.10 0.09
0.216
0.20 0.16
0.317
* Significantly different from zero at the 0.05 probability level 
** Significantly different from zero at the 0.01 probability level
Negative and significantly different from zero at the 0.05 probability level
98
Table 4.4 Cont
Cross
Across Environments 
Mean SCA
Ibadan 1999 
Mean SCA
I30572 X 14(2)1425 1.69 -0.22 2.36 -0.09
I30572 X 160142 1.79 -0.11 2.62 0.19
I30572 X 190257 1.48 -0 43* 11 .98 -0.15
130572 X TME1 1.95 0.05 ..3.08 0.49*
130572 X TME10 1.77 -0.13 2.63 -0.20
130572 X TME11 1.74 -0.17 2.37 -0.21
130572 X TME117 2.19 0.29* 3.45 0.21
130572 X TME12 2.08 0.18 2.66 0.10
130572 X TME14 1.91 0.04 2.58 -0.24
130572 X TME2 2.34 0.44* 3.63 -0.11
130572 X TME31 2.15 0.24* 3.34 -0.10
130572 X TME4 1.93 0.02 2.97 0.22
130572 X TME41 1.97 0.06 3.03 0.03
130572 X TME5 1.84 -0.06 2.63 -0.14
130572 X TME6 1.73 -0.17 2.66 -0.11
130572 X TME7 1.86 -0.04 2.68 0.01
130572 X TME8 1.86 -0.04 2.61 0.14
130572 X TME9 1.96 0.05 2.44 -0.05
SE 0.10 0.09 0.20 0.16
LSD 0.216 0.317
* Significantly different from zero at the 0.05 probability level 
** Significantly different from zero at the 0.01 probability level
Negative and significantly different from zero at the 0.05 probability level
99
Significant and negative SCA was detected for TMS 130001 X TMS 160142 in Ibadan 
1999 and across environments, for TMS 130001 X TME1 in Ibadan 1999 and also for 
TMS I30555 XTME12 across environments.
Significant and positive SCA effects were estimated for all the three crosses involving 
TME2, TMS 130001 X TME31, TMS I30555 X TME5, TMS I30572 X TME117 and 
TMS I30572 X TME31 across environments, and TMS 130001 X TME14, TMS 
I30555 X TME11 and TMS I30572XTME1 in Ibadan 1999.
4.1.4 Combining ability and maternal effects in the 4X4 Diallel mating scheme
The parental means, GCA and maternal effects of the four parents used in 
environments with significant variation for these effects are given in (Table 4.5). 
Significant and negative GCA effects were detected for the clones TMS I30572 and 
clone TMS 130001 in Mokwa 1998 and also for TMS I30572 in Ibadan 1998. 
Significant and positive GCA effects were also detected for the clones TMS I30555 in 
Mokwa 1998 and 58308 in Ibadan 1998. The maternal effect of 58308 was also 
significant and positive in the Mokwa 1998 and Ibadan 1998 environments.
In Mokwa 1998 significant and negative SCA effects were detected for the crosses 
TMS 130001 X TMS I30555, TMS I30555 X TMS I30572 and in Ibadan 1999 
environment, for the crosses TMS 130001 X 58308 and TMS I30555 X TMS I30572 
(Table 4.6). The self crosses TMS 130001 X TMS 130001, TMS I30555 x TMS 
I30555, and 58308 X 58308 had significant and positive SCA effects in Mokwa 1998 
environment, and in Ibadan 1999, TMS 130001 X TMS 130001, 58308 X 58308 and 
I30572 X 58308 were also significant and positive.
100
Table 4.5 Mean, GCA and maternal (MAT) effects for CMD symptom severity of parents
in 4X4 diallel crosses in environments with significant variation for their effects
Mokwa 1998 Ibadan 1998 Ibadan 1999*
C lone Mean GCA MAT Mean GCA MAT Mean GCA MAT
130001 1.00 -0.18* -0.02 1.01 -0.07 -0.13 1.34 0.00
I30555 1.95 0.21* -0.03 2.01 0.10 -0.12 2.35 0.23
130572 1.10 -0.10* -0.03 1.63 -0.18* 0.09 1.95 -0.03
58308 X 0.07 0.09* X 0.16* 0.17* 1.13 -0.20
Mean 1.35 1.55 1.69
SE 0.25 0.20 0.34
LSD 0.086 0.086 0.136 0.136 0.245
LSD gi-gj 0.156 0.237 0.411
LSD mj_mj 0.156 0.237
x Not estimated due to missing plot
* Maternal e ffect not s ign ifican t in Ibadan 1999 environm ent
* S ign ificantly d iffe ren t from  ze ro  at the 0.05 probability  level 
** S ignificantly d iffe ren t from  zero a t the 0.01 probability level
101
Table 4.6 Mean CMD severity scores and SCA effects for CMD symptom severity of
crosses in 4X4 diallel in environments with significant variation for their effects
Cross M o kw a 1998 Ibadan 1999
Mean SCA Mean SCA
130001 X  130001 1.23 0.18* 3.14 0.64*
130001 X  I30555 1.24 -0.12* 2.40 -0.08
130001 X I30572 1.15 0.03 2.51 -0.02
130001 X 58308 1.18 -0.09 2.08 -0.54*
I30555 X  I30555 2.16 0.34* 3.07 0.13
I30555 X I30572 1.35 -0.18* 2.33 -0.34*
I30555 X  58308 1.42 -0.04 2.80 0.29
I30572 X I30572 1.24 0.04 2.43 0.01
I30572 X  58308 1.36 0.11 2.66 0.35*
58308 X  58308 1.55 0.29* 2.00 0.35*
Mean 1.40 2.50
SE 0.11 0.38
LSD sii 0.156 0.410
LSD „ 0.116 0.308
LSD Sjj-Sjj 0.208 0.548
LSD sii-sij 0.224 0.612
LSD sii-sjk 0.180 0.474
LSD stj-sik 0.180 0.474
LSD sij-skl 0.146 0.388
* S ignificantly d iffe ren t from  ze ro  at the 0.05 probability  level 
** S ignificantly d iffe rent from  zero at the 0.01 probability  level
Negative and s ign ificantly d iffe rent from  zero at the  0.05 probability  level
102
The significant reciprocal effect in Ibadan 1998 was due to the negative and 
significant effect of three reciprocal crosses 58308 X 130001, I30572 X I30555 and 
58308 XI30555.
4.1.5 Heterosis for resistance to CMD in F! crosses
In the NCD II experiment, average values for mid-parent and high-parent heterosis 
were generally positive and indicated that the F-i crosses were generally more 
susceptible than their mid or high parents except in the Mokwa 1998 environment 
where the Ft differed from the mid-parent value by -4.95%.
Significant and negative heterosis was estimated in only two crosses involving clone 
TMS I30555 as the female parent, and TME117 and TME41 as the male parents. 
The most heterotic cross, which contributed significantly to the average midparent 
heterosis, was TMS I30555 X TME41. It had negative and significant mid parent 
heterosis of -31.94% across environments and in Mokwa 1998, both mid-parent and 
high-parent heterosis of -48.44% and -40.51% were negative and significant. The 
next most heterotic cross was TMS I30555 X TME117, which also had negative 
heterosis in all environments. Its mid-parent heterosis of -50.00% and high-parent 
heterosis o f-41.03% were significant in Mokwa 1998.
Generally, the crosses in the diallel experiment had positive heterotic effects for both 
mid-parent and high-parent heterosis and were generally more susceptible than their 
mid-parent or high-parents. Negative heterosis was detected only in the self cross 
TMS I30572 X TMS I30572 in Ibadan 1998 and for most crosses in the Mokwa 
environment, however, none of these negative heterotic effects were significant.
103
4.2 Genetic relationships among the various sources of resistance to CMD
4.2.1 Relationships among sources of resistance based on area under the 
disease progress curve (AUDPC) for CMD
Estimated AUDPC values of the cassava clones used as parents and checks and 
their respective rankings in each and across environments for CMD resistance 
responses are shown in Table 4.7.
Clone 91/02324 and TME9 had the smallest AUDPC values for the shoot-tip and 
whole plant severity and were ranked 1 (i.e. best resistant clones) across 
environments. In Ibadan 1998 and 1999, TME9 consistently had a rank of 1 for shoot 
tip severity but ranked 4 for whole plant in Mokwa 1998. The clone 91/02324 also 
ranked 3 for whole plant severity in the Ibadan 1999 environment.
The next best clone was the resistant improved line TMS 130001, which ranked 3 
across environments for both shoot tip and whole plant, and also ranked 1 for shoot 
tip severity in the individual environments and 4 for whole plant severity in Ibadan 
1998. In the Ibadan 1998 environment, 160142 also ranked 1 for shoot tip and whole 
plant severity and in Mokwa 1998, for shoot tip severity.
In the Ibadan 1998 environment, the resistant landraces TME5, TME7, TME8 TME11 
TME12 and TME14 ranked 1 for shoot tip severity, while in Ibadan 1999, resistant 
landraces, TME4 and TME5 ranked 1 for shoot tip severity in Ibadan.
Based on their AUDPC values and ranks across environments and in Ibadan 1998 
and 1999, the results further showed that the resistant landraces were generally 
more resistant than improved clones TMS I30572 and TMS 14(2)1425. However, in 
Mokwa 1998, TMS I30572 had lower ranks for both shoot tip and whole plant 
severity, indicating that it was more resistant than resistant landraces, TME5, TME7, 
TME8, TME9, TME12 and TME14.
104
Table 4.7 Area under the disease progress curve (AUDPC) of cassava clones based on 
mean CMD severity scores and their ranks for CMD resistance.
Clone Across
environments
Mokwa1998 Ibadan 1998 Ibadan 1999
AUDPC Rank AUDPC
Shoot Tin Severitv 
Rank AUDPC Rank AUDPC Rank
58308 X X X X X X 12.00 1
91/02324 20.00 1 X X 20.00 1 12.00 1
TMS 130001 22.60 3 12.00 1 20.00 1 12.00 1
TMS I30555 33.40 17 27.73 17 20.00 1 32.80 19
TMS I30572 29.31 15 13.28 7 28.23 16 18.50 15
TMS 14(2)1425 31.79 16 18.97 15 30.46 17 26.40 17
TMS 160142 27.91 13 12.00 1 20.00 1 22.30 16
TMS 190257 24.75 8 16.43 13 23.88 13 16.00 12
TME1 26.49 11 17.17 14 23.52 12 16.30 13
TME2 39.88 18 40.47 19 35.64 18 30.80 18
TME4 25.75 10 12.00 1 22.52 11 12.00 1
TME5 24.66 7 15.59 10 20.00- 1 12.00 1
TME6 26.87 12 12.00 1 26.68 14 14.30 9
TME7 23.3 4 14.55 8 20.00 1 16.80 14
TME8 23.36 5 15.45 9 20.00 1 14.10 8
TME9 20.00 1 12.00 1 20.00 1 12.00 1
TME10 45.88 19 26.17 16 42.32 20 41.70 20
TME11 28.03 14 12.00 1 27.14 15 15.20 11
TME12 25.32 9 16.30 12 20.00 1 14.30 9
TME14 23.65 6 16.03 11 20.00 1 12.00 1
TME31 51.20 20 50.61 20 40.78 19 47.50 21
TME41 53.27 21 38.10 18 44.56 21 56.90 22
TME117 68.05 22 51.29 21 65.09 22 63.32 23
Whole Plant Severitv
AUDPC Rank AUDPC Rank AUDPC Rank AUDPC Rank
58308 X X X X X X 12.00 1
91/02324 20.00 1 X X 20.00 1 14.71 3
TMS 130001 24.51 3 12.00 1 23.60 4 18.09 4
TMS 130555 38.52 17 33.79 16 35.12 16 49.63 19
TMS 130572 33.09 14 15.11 5 35.65 17 24.42 11
TMS 14(2)1425 35.59 16 20.75 15 33.82 15 43.73 17
TMS 160142 33.11 15 12.00 1 27.78 10 38.62 16
TMS 190257 27.37 6 19.12 10 27.53 9 23.92 9
TME1 31.56 13 19.10 9 29.66 13 26.47 13
TME2 45.19 18 47.98 19 42.40 18 45.29 18
TME4 28.95 9 12.00 1 26.19 8 23.34 8
TME5 27.78 8 16.77 8 24.43 7 21.93 7
TME6 30.39 10 19.85 11 28.73 11 24.30 10
TME7 27.05 5 16.61 7 23.60 4 20.52 5
TME8 27.64 7 19.93 12 20.00 1 28.32 15
TME9 20.00 1 13.81 4 20.00 1 12.00 1
TME10 51.66 19 38.99 17 50.81 20 55.42 20
TME11 31.41 12 20.38 14 32.27 14 21.27 6
TME12 31.19 11 20.18 13 29.44 12 25.98 12
TME14 25.91 4 16.33 6 24.30 6 27.07 14
TME31 54.72 20 54.48 21 45.78 19 57.5 21
TME41 57.28 21 44.93 18 55.47 21 70.48 22
TME117 70.38 22 52.73 20 70.58 22 72.52 23
x Not estimated due to missing plot
105
Table 4.8 Spearman’s rank correlation coefficient for AUDPC values for shoot-tip and 
whole plant CMD symptom severity
Across environm ents Mokwa 1998 Ibadan 1998 Ibadan 1999
Across Env
Shoot-tip
0.678** 0.835** 0.886**
M okwa 1998 0.678** 0.573** 0.744**
Ibadan 1998 0.835** 0.573** 0.727**
Ibadan 1999 0.886** 0.744** 0.727**
Across Env
W hole  P lant 
0 .754** 0.951** 0.880**
Mokwa 1998 0.754** 0.757** 0 .741**
Ibadan 1998 0.951** 0.757** 0 .792**
Ibadan 1999 0.880** 0.741** 0.792**
106
As was expected, the susceptible improved clone (TMS 130555) and the susceptible 
landraces (TME2, TME10, TME31, TME41 and TME117) had the largest AUDPC 
values and had ranks higher than the resistant clones. With the exception of whole 
plant severity in Mokwa 1998, TME117 was the least resistant and most susceptible 
clone. TMS I30555 was the least susceptible followed by TME2 or TME10 depending 
on the environment.
The rank correlation coefficients among the AUDPC values showed significant and 
positive linear associations among the AUDPC values in the different environments 
for both shoot-tip and whole plant CMD symptom severity (Table 4.8). The correlation 
coefficient (r) ranged between 0.57 to 0.89 for shoot tip CMD severity and between 
0.74 to 0.95 for whole plant CMD severity.
4.2.2. Relationships among sources of resistance based on principal 
component analysis (PCA) of CMD responses
Eigenvalues (proportion of the total variance) and eigenvectors for the first three 
principal components (PC1, PC2 and PC3) for the CMD resistance responses among 
the 23 parental and check clones are presented in Table 4.9. The eigenvalues for the 
first three principal components accounted for 78.87% of the total variation.
107
Table 4.9 Eigenvectors and eigenvalues as a proportion of total variance of the first 
three principal components for the CMD resistance responses of the 23 parental and 
check clones
Environment CMD Response PC1 PC2 PC3
Mokwa 1998 MST6WK 0.218 -0.058 0.238
MCMD6WK 0.238 0.003 0.186
MST12WK 0.232 -0.075 0.132
MCMD12WK 0.228 -0.058 0.176
ST6WKI 0.082 0.389 0.205
CMD6WKI 0.163 0.047 0.304.
ST12WKI 0.231 -0.047 0.142
CMD12WKI 0.201 -0.118 -0.032
Ibadan 1998 MST6WK 0.229 -0.006 -0.198
MCMD6WK 0.227 -0.002 -0.230
MST12WK 0.228 -0.008 -0.196
MCMD12WK 0.225 0.001 -0.236
MST20WK 0.218 -0.089 -0.013
MCMD20WK 0.238 -0.052 -0.015
ST6WKI 0.048 0.387 -0.230
CMD6WKI 0.099 0.339 -0.185
ST12WKI 0.041 0.411 -0.239
CMD12WKI 0.106 0.323 -0.169
MST20WKI 0.101 0.320 0.126
MCMD20WKI -0.007 0.317 0.362
Ibadan 1999 MST6WK 0.239 -0.067 -0.076
MCMD6WK 0.237 -0.030 0.014
MST12WK 0.237 -0.082 0.087
MCMD12WK 0.231 -0.043 0.099
ST6WKI 0.195 -0.053 -0.146
CMD6WKI -0.123 0.153 0.279
ST12WKI 0.151 0.161 0.252
CMD12WKI -0.138 0.024 0.111
Eigenvalues (%) 57.33 13.78 7.76
Mean shoot-tip CMD severity at 6 WAP 
Mean shoot-tip CMD severity at 12 WAP 
Mean whole plant CMD severity at 6 WAP 
Mean whole plant CMD severity at 12 WAP 
Mean shoot-tip CMD severity at 20 WAP 
Mean whole plant CMD severity at 20 WAP 
Mean shoot-tip CMD incidence at 6 WAP 
Mean shoot-tip CMD incidence at 12 WAP 
Mean whole plant CMD incidence at 6 WAP 
Mean whole plant CMD incidence at 12 WAP
MST6WK
MST12WK
MCMD6WK
MCMD12WK
MST20WK
MCMD20WK
ST6WKI
ST12WKI
CMD6WKI
CMD12WKI
108
The first PC (PC1), which accounted for 57.33% of the total variation gave higher 
weights to severity of shoot tip and whole plant at 6 and 12 WAP in Mokwa, 6, 12 and 
20 WAP in Ibadan 1998, and 6 and 12 WAP in Ibadan 1999. In the Mokwa 1998 
environment, PC1 also gave a high weight for shoot tip incidence at 12 WAP. The 
second principal component (PC2), which accounted for 14% of the total variation, 
gave higher weights to shoot tip and whole plant incidence at 6, 12 and 20 WAP in 
the Ibadan 1998 environment. The third principal component (PC3) also gave high 
weightings to shoot tip severity at 6 WAP and whole plant incidence at 6 WAP in 
Mokwa 1998. In the Ibadan 1998 environment, whole plant severity at 6 and 12 
WAP, shoot tip incidence at 6 and 12 WAP, and whole plant incidence at 20 WAP 
also had high weightings in PC3, but PC3 accounted for only 8% of the total 
variation.
The scatter-plot of the first versus second principal component scores for the CMD 
response of the various sources of resistance in the three environments is presented 
in Fig 4.1. Since 58308 and 91/02324 were not present in all environments they were 
excluded from the analysis. The results showed that, generally, the resistant clones 
grouped together and were different from the susceptible. The first resistant group 
was made up of the resistant landraces TME8, TME9, TMS 130001, TME7 and 
TME12. TMS I90257clustered together with TME1 and TME6. The resistant 
landraces TME5 and TME14 formed another group, and TME11 clustered together 
with TMS 160142 and TMS I30572. The resistant landraces TME4 and the improved 
clone TMS 14(2)1425 formed single groups. TMS I4(2)1425’s group was close to the 
largest susceptible group made up of clones TME10 only, TME31 and TME2. The 
susceptible landrace TME117 formed a single group away from all the other 
susceptible groups. TME41 and TMS I30555 also formed single susceptible groups.
109
CO
CM<O
PCA 1 (57 .33% )
Fig 4.1 Plot of first principal component versus second principal component analysis 
of parents and checks across environments 
Alphabeths represent clones;
C=TMS 130001,
G=TMS 160142,
K=TME11,
0=TME2,
S=TME5,
W=TME9
D=TMS I30555, 
H=TMSI90257, 
L=TME117, 
P=TME31, 
T=TME6,
E=TMS I30572,
l=TME1,
M=TME12,
Q=TME4,
U=TME7,
F=I4(2)1425,
J=TME10,
N=TME14,
R=TME41,
V=TME8,
110
4.2.3 Complementarity of genes for resistance to CMD among various 
sources of resistance
4.2.3.1 Frequency distribution of CMD severity scores in F: progenies of various 
sources of resistance
The frequency distributions of the mean disease severity scores of the F-i progenies 
in crosses between susceptible parents, resistant parents and between resistant and 
susceptible parents are presented in Tables 4.10, 4.11 and 4.12 respectively. With 
scores of 1 and 2 taken as resistant, and scores of 3 to 5 taken as susceptible, the 
proportion of resistant to susceptible progenies varied from cross to cross and across 
environments but higher levels of resistance were generally obtained in the Mokwa 
1998 environment. In all the three environments, there were generally fewer plants in 
the extremely susceptible class with the disease severity score of 5 for the various 
crosses.
The susceptible by susceptible crosses TMS I30555 X TME41 and the self cross 
TMS I30555 XTMS I30555 had about 50% of their progenies in the resistant classes 
of 1 and 2 in Ibadan 1998 and 1999. The other crosses in this category had over 
60% of their progenies being susceptible (Table 4.10).
Susceptible progenies were detected in all the resistant by resistant crosses in at 
least one environment (Table 4.11). The best resistant by resistant cross was TMS 
I30572 X TMS I90257 which had relatively fewer susceptible progenies across 
environments (9%), in Ibadan 1998 (10.8%), in Ibadan 1999 (24%) and in Mokwa 
1998 (0%). The worst resistant by resistant cross was TMS I30572 X TME12 with 
32.9% of its progenies being susceptible across environments, 14.3% in Mokwa 
1998, 37.2% in Ibadan 1988, and 56.2% in Ibadan 1999.
111
Table 4.10. Frequency distribution of Fi disease severity scores (based on mean disease severity scores of progenies) in an environment and mean
number of progenies (N) in crosses involving susceptible parents.
Across environments Mokwa 1998 Ibadan 1998 Ibadan 1999
Cross
CMD severity scores N CMD severity scores N CMD severity scores N CMD severity scores N
1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5
I30555 X I30555 2 10 4 1 0 17 3 9 4 1 0 17 1 13 4 2 0 20 1 5 2 6 2 16
I30555 X TME2 24 3 10 18 1 56 49 7 10 4 0 70 22 2 15 10 1 50 0 1 6 39 3 49
I30555XTME10 47 7 12 11 2 79 100 9 15 3 0 127 36 3 8 16 2 65 5 8 14 13 4 44
I30555 X TME31 20 4 4 8 0 36 41 2 3 2 0 48 18 6 4' 2 0 30 2 4 6 19 0 31
I30555 X TME41 48 9 5 12 1 75 81 4 3 2 0 90 52 6 2 7 1 68 10 17 9 26 2 64
I30555XTME117 10 2 3 3 1 19 21 3 1 0 0 25 10 1 3 2 1 17 0 3 4 8 1 16
112
Table 4.11 Frequency distribution of Fi disease severity scores (based on mean disease severity scores of progenies) in an environment and mean
number of progenies (N) in crosses involving resistant parents.
Across environments Mokwa 1998 Ibadan 1998 Ibadan 1999
Cross CUD severity scores N CMD severity scores N CMD seventy scores N CMD severity scores N
1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5
130001 X 130001 8 2 1 1 0 12 9 2 1 0 0 12 5 1 2 2 0 10 1 1 2 3 0 7
130001 XI30572 29 9 1 0 0 39 35 6 0 0 0 41 18 14 4 0 0 36 4 10 6 6 0 26
130001 X 158308 6 3 0 0 0 9 7 2 0 0 0 9 3 6 1 0 0 10 2 4 1 1 0 8
130001 X 14(2)1425 5 5 2 1 1 14 10 5 1 0 0 16 1 5 3 2 1 12 3 4 3 2 1 13
130001 X 160142 21 6 6 3 0 36 40 3 2 0 0 45 12 7 10 6 0 35 10 7 7 3 0 27
130001X 190257 6 4 2 1 0 13 12 2 1 0 0 15 3 6 1 2 0 12 2 4 4 2 0 12
130001 X  TME1 3 3 2 1 0 9 6 3 1 0 0 10 2 3 3 1 0 9 2 3 2 1 0 8
130001 X  TME4 12 3 4 5 0 24 23 2 6 2 0 33 10 3 4 4 1 22 4 4 2 9 0 19
130001 X TME5 15 9 6 4 1 35 28 8 3 1 1 41 13 7 9 4 1 34 3 11 5 8 0 27
130001 X TME6 9 4 4 2 0 19 16 5 2 1 0 24 7 4 5 2 0 18 3 2 5 3 1 14
130001 X  TME7 15 7 6 7 0 35 27 8 3 2 0 40 12 7 10 5 0 34 6 5 5 13 1 30
130001 X TME8 6 8 1 1 0 16 12 9 2 0 0 23 4 7 2 1 0 14 2 7 0 3 1 13
130001 X TME9 30 8 9 9 1 57 50 8 8 1 0 67 27 7 13 6 0 53 13 10 5 20 2 50
130001 XTME11 8 4 2 3 0 17 12 4 2 1 0 19 7 5 2 2 0 16 5 3 1 5 0 14
130001 X TME12 8 2 1 2 1 14 15 2 0 1 0 18 6 3 1 1 1 12 3 2 1 3 1 10
130001 X TME14 4 2 1 2 0 9 8 1 1 1 0 11 3 2 3 1 0 9 1 2 0 4 1 8
I30572 XI30001 21 6 2 1 0 30 25 5 0 0 0 30 13 9 7 2 0 31 7 7 3 7 0 24
130572 X 130572 10 2 1 0 0 13 11 3 0 0 0 14 7 1 3 0 0 11 3 1 0 3 0 7
I30572X 58308 35 20 7 2 0 64 43 19 2 0 0 64 18 23 16 7 0 64 10 14 6 17 1 48
130572X14(2)1425 33 5 4 3 0 45 60 5 2 0 0 67 26 3 5 3 0 37 13 6 5 5 0 29
130572X160142 15 3 3 3 0 24 28 1 0 0 0 29 13 1 4 2 0 20 4 6 5 6 0 21
I30572 X I90257 32 7 2 2 0 43 53 4 0 0 0 57 28 5 2 2 0 37 14 11 3 5 0 33
113
Table 4.11. continued
Across environments M okwa1998 Ibadan 1998 Ibadan 1999
Cross
CMD severity scores N CMD severity scores N CMD severity scores N CMD severitv scores N
1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5
130572 X TME1 6 1 0 2 0 9 10 2 1 0 0 13 5 1 0 1 0 7 2 0 0 4 1 7
130572 X TME4 44 7 5 13 1 70 76 6 4 0 0 86 44 5 7 8 0 64 13 11 5 31 2 62
I30572 X TME5 37 6 5 8 0 56 61 2 5 1 0 69 37 5 6 8 0 56 12 11 3 16 1 43
130572 X TME6 32 5 4 6 0 47 51 3 3 0 0 57 33 2 5 2 0 42 11 9 4 15 1 39
130572 X TME7 20 3 3 5 0 31 32 4 4 0 0 40 20 2 3 4 0 29 9 3 3 11 1 27
130572 X TME8 47 10 3 11 0 71 91 7 1 2 0 101 34 11 4 11 0 60 15 12 3 21 1 52
130572 X TME9 157 21 37 44 2 261 237 33 54 10 1 335 136 15 34 39 3 227 97 14 22 84 1 218
130572 XTME11 52 20 3 10 0 85 84 25 2 7 0 118 52 14 2 4 0 72 20 21 6 19 0 66
I30572 X TME12 45 12 10 17 1 85 69 21 10 5 0 105 42 7 10 18 1 78 24 8 11 28 2 73
I30572 X TME14 41 15 7 12 1 76 66 20 5 3 0 94 42 8 7 12 1 70 16 16 8 21 1 62
58308 X 130001 4 3 2 0 0 9 5 2 0 0 0 7 2 4 6 0 0 12 4 1 1 0 0 6
58308 X I30572 18 16 6 2 0 42 22 18 4 1 0 45 11 11 10 4 1 37 6 10 5 9 0 30
58308 X 58308 2 2 1 0 0 5 3 2 1 0 0 6 1 3 1 1 0 6 1 3 0 1 0 5
114
Table 4.12 Frequency distribution of F, disease severity scores (based on mean disease severity scores of progenies) in an environment and mean
number of progenies (N) in crosses involving resistant and susceptible parents.
Across environments Mokwa 1998 Ibadan 1998 Ibadan 1999
Cross
CMD severity scores N CMD severity scores N CMD severity scores N CMD severity scores N
1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5
130001 X 130555 28 12 4 1 0 45 35 10 1 0 0 46 13 17 11 2 0 43 7 15 7 7 0 36
130001 X TME2 9 4 9 7 1 30 22 5 9 1 0 37 6 6 11 6 1 30 0 0 6 15 1 22
130001 XTME10 13 5 8 5 0 31 26 4 3 1 0 34 10 5 11 3 1 30 3 6 9 10 0 28
130001 X TME31 5 1 3 2 0 12 10 3 1 0 0 14 5 1 4 0 1 11 0 0 3 7 0 10
130001 X TME41 16 5 4 7 0 32 30 9 1 0 0 40 13 4 7 4 0 28 4 3 4 16 1 28
130001 XTM E117 12 6 4 4 0 25 22 9 1 0 0 32 11 6 5 4 0 26 3 2 5 8 0 18
I30572X 130555 19 9 5 1 0 34 24 9 1 0 0 34 9 9 12 2 0 32 4 14 2 6 0 26
I30572 X TME2 18 3 7 8 0 36 40 5 6 0 0 51 15 1 8 5 1 30 0 2 6 19 0 27
[30572 XTME10 37 5 9 6 0 57 55 7 2 0 0 64 41 2 10 3 0 56 15 6 16 14 1 52
130572 X TME31 37 7 9 14 1 68 64 9 4 1 0 78 38 10 9 9 1 67 9 2 13 32 3 59
I30572 X TME41 49 3 4 14 1 71 94 7 3 1 1 106 37 1 7 8 0 53 16 2 3 33 1 55
I30572 X TME117 53 15 14 25 1 108 92 20 5 1 0 118 63 9 18 15 1 106 3 15 19 58 3 98
I30555 X 130001 24 12 6 1 0 43 30 11 3 0 0 44 12 14 12 4 1 43 5 10 5 17 1 38
I30555X 130572 21 10 3 0 0 34 24 12 1 0 0 37 15 6 7 0 0 28 8 4 2 7 0 21
I30555X 58308 13 8 2 1 0 24 15 8 1 0 0 24 9 8 3 3 0 23 2 7 0 6 1 16
115
Tab le  4 .12  C on tin ued
Across environments Mokwa 1998 Ibadan 1998 Ibadan 1999
Cross
CMD severity scores N CMD severity scores N CMD severity scores N CMD severity scores N
1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5
I30555 X 14(2)1425 6 4 1 2 0 13 12 5 1 0 0 18 4 3 1 2 0 10 2 3 1 4 0 10
130555X160142 25 5 3 4 0 37 47 4 1 1 0 53 23 3 1 3 0 30 5 8 7 9 0 29
I30555 X I90257 15 8 1 1 0 25 35 6 0 0 0 41 5 10 2 2 0 19 5 8 2 2 0 17
130555 X TME1 25 10 5 4 0 44 45 5 3 0 0 53 24 10 6 2 0 42 5 16 5 10 0 36
I30555 X TME4 34 7 3 7 1 52 56 5 3 1 0 65 31 5 4 5 0 45 14 10 2 14 2 42
I30555 X TME5 19 6 7 8 1 41 37 4 6 1 0 48 18 7 7 6 2 40 3 8 8 17 1 37
I30555 X TME6 53 9 8 14 2 86 86 9 6 1 0 102 60 2 7 10 2 81 13 15 10 32 3 73
130555 XTME7 49 11 5 10 2 77 84 11 5 3 1 104 39 12 4 4 2 61 25 11 5 23 3 67
I30555 X TME8 43 16 4 7 0 70 77 13 3 1 0 94 37 15 4 5 0 61 14 19 4 16 0 53
130555 XTME9 27 3 4 7 0 41 40 3 4 3 0 50 27 1 1 7 0 36 14 5 6 10 1 36
I30555 X TME11 29 13 5 9 1 57 58 6 7 2 0 73 24 17 4 4 0 49 4 15 4 20 2 45
I30555 X TME12 40 5 4 6 0 55 72 4 3 1 0 80 36 2 4 4 0 46 13 10 6 14 0 43
I30555 X TME14 29 7 3 5 0 44 68 7 6 3 0 84 18 3 3 5 0 29 2 11 1 8 1 23
58308 X 130555 3 6 3 1 0 13 4 7 2 0 0 13 2 3 5 2 0 12 0 4 0 3 0 7
116
The crosses involving resistant and susceptible parents had progenies falling into all 
classes in the Ibadan 1998 and Ibadan 1999 environments (Table 4.12). The crosses 
involving the resistant parent TMS 130001, however, had fewer plants in the resistant 
classes of 1 and 2 in Ibadan 1999. The crosses TMS 130001 X TME2 and TMS 
130001 X TME31 had no resistant progenies in that environment. Another poor cross 
in Ibadan 1999 was TMS I30572 X TME2 with only 4 out of 54 progenies having the 
resistant score of 2. The best cross between a susceptible and resistant parent was 
the cross TMS I30555 X TMS I90527, with 96%, 100%, 78% and 76% of its progeny 
being resistant across environments, in Mokwa 1998, Ibadan 1998, and Ibadan 1999 
respectively.
4.2.3.2 Transgressive segregants in Fi progenies of various sources of resistance
Positive transgressive segregant (PTS), individual progenies with a mean CMD 
severity scores at least one disease severity score lower than the better parent, were 
desirable for resistance and are presented for the different cross combinations in 
Table 4.13. PTS were detected in six susceptible by susceptible crosses, 14 resistant 
by susceptible crosses and eight resistant by resistant crosses. The highest 
frequencies of PTS across environments were detected in six susceptible by 
susceptible crosses, which ranged from 20.83% in the self, I30555XI30555 to 72% in 
I30555XTME41. Similarly, the lowest frequency of TS was detected in the resistant 
by resistant cross TMS I30572 XTMS 160142 (6.17%), and I30572XTME14 (8.83%).
117
Table 4.13 Frequency of positive transgressive in percentages (PTS) with disease 
severity scores at least one score better than the better parent in a cross in three 
environments
Cross Across environments Mokwa 1998 Ibadan 1998 Ibadan 1999
Sus x Sus PTS N PTS N PTS N PTS N
I30555 X 130555 20.83 18 15.50 17 5.50 20 41.50 16
I30555 X TME2 42.5 56 80.5 70 45.0 50 2.0 49
I30555 X TME10 56.3 79 80.0 127 57.5 65 31.5 44
I30555 X TME31 57.8 36 91.5 48 63.5 30 18.5 31
I30555 X TME41 72.0 75 95.0 90 77.0 68 44.0 64
I30555 X TME117 59.8 19 98.0 25 62.0 17 19.5 16
Sus x Res
I30555X 130572 32.33 34 55.00 28 42.00 21
I30555X 14(2)1425 33.00 13 49.50 10 49.50 10
130555 X 160142 15.17 37 45.50 29
I30555 X I90257 18.83 25 26.50 19 30.00 17
I30555 X TME1 23.33 44 58.00 42 12.00 36
130555 X TME8 8.50 70 25.50 53
130555 XTME11 16.33 57 49.00 49
I30555 XTME14 3.00 44 9.00 23
Res x Sus
I30572X 130555 15.2 34 29.0 32 16.5 26
130572 X TME2 18.2 36 54.5 30
130572 XTME10 34.0 57 74.0 56 28.0 52
130572 X TME31 23.7 68 57.0 67 14.0 59
I30572 X TME41 33.3 71 71.0 53 29.0 55
I30572 X TME117 21.0 108 60.0 106 3.0 98
Res x Res
130572XI30572 36.50 13 67.00 11 42.50 7
I30572 X 14(2)1425 36.17 45 71.50 37 37.00 29
130572X 160142 6.17 24 18.50 21
130572 X I90257 39.17 43 76.00 37 41.50 33
I30572 X TME1 36.50 9 76.50 7 33.00 7
I30572 X TME8 9.83 71 29.50 52
I30572 X TME11 23.83 85 71.50 72
I30572 X TME14 8.83 76 26.50 62
N number o f p rogen ies in an environm ent and across environm ents 
Sus Susceptib le  
Res Resistant
118
The CMH test based on the mean distribution of F-, progeny disease severity scores 
of the sources of resistance, revealed significant differences among crosses between 
TMS 14(2)1425 and the resistant landraces TME12, TME14, and TME9 as males and 
the resistant clone TMS I30572 as the female (Table 4.14). Significant differences 
were also detected among crosses between TMS 160142 and TME5 as the male 
parents and TMS I30555. The resistant improved clone TMS I90257 exhibited 
significant difference from the resistant landraces, TME5 when these were crossed 
with TMS I30555 and with TME9, TME12 and TME14 when these were crossed with 
TMS I30572. TME12 and TME11 were significantly different based on their crosses 
with TMS I30572, while TME5, TME8, TME12 and TME14 were significantly different 
based on their crosses with TMS I30555.
The results showed that susceptible clone TMS I30555 was significantly different 
from resistant clones TMS 130001, 58308 and TMS I30572 when these were crossed 
with TMS I30555. The susceptible clones TME2, TME31 and TME117 were also 
different from the resistant clones TMS 14(2)1425 and TMS I90257 when these were 
crossed with TMS I30572 (Table 4.15). The results further showed that TME2 was 
significantly different from TMS 160142 based on their crosses with TMS 130001 and 
TMS I30555 and different from TMS I90257, TME1, TME4, TME6, TME7, TME8, 
TME9 and TME11 from their crosses with TMS I30555. Significant difference 
between TME2 and TME11 was also detected when they were crossed with TMS 
I30572. The susceptible clone TME117 was also significantly different from TME5, 
TME6, TME8 and TME11 when these were crossed with TMS I30572.
The results further showed that TME2 was significantly different from TME10 and 
TME41 from their crosses with TMS I30555 and that TME117 and TME10 were 
significantly different from their crosses with TMS I30572 (Table 4.16).
4.2.3.3 Cochran-Mantel-Haenzel (CMH) statistic test of association
119
Table 4. 14 Chi-square values for differences in the mean distribution of disease 
severity scores of Ft progenies in crosses involving resistant male parents when 
crossed with each of the female parents 130001 resistant to CMD, I30555 susceptible to 
CMD and I30572 resistant to CMD
Resistant male 58308
Female parent 
130001 I30555 I30572
130001 vs 58308 0.00 0.49 0.00 1.18
130001 vs  I30572 0.01 1.92 0.83 0.27
I30572 vs 58308 0.00 0.08 0.63 1.76
14(2)1425 vs 160142 1.31 0.79 1.09
14(2)1425 vs I90257 0.48 1.68 0.26
14(2)1425 vs TME1 0.00 0.37 0.65
14(2)1425 vs TME11 0.11 0.00 0.89
14(2)1425 vs TME12 0.08 1.23 6.11*
14(2)1425 vs TM E14 0.00 0.75 4 .06*
14(2)1425 vs  TME4 0.02 0.29 2.83
14(2)1425 vs TME5 0.05 0.39 1.17
14(2)1425 vs TME6 0.24 0.02 0.66
14(2)1425 vs TME7 0.00 0.20 1.40
14(2)1425 vs TME8 0.76 0.86 1.02
14(2)1425 vs TME9 0.15 0.15 4 .62*
160142 vs 190257 0.09 0.18 2.25
160142 vs TME1 0.89 0.22 0.00
160142 vs TME11 0.63 1.85 0.15
160142 vs TME12 0.49 0.07 0.93
160142 vs TME14 0.82 0.00 0.34
160142 vs TME4 1.27 0.21 0.14
160142 vs TME5 1.39 4.09* 0.02
160142 vs TME6 0.45 1.14 0.11
160142 vs TME7 2.18 0.40 0.01
160142 vs TME8 0.05 0.01 0.05
160142 vs TME9 1.04 0.39 0.34
I90257 vs TME1 0.37 0.81 1.36
I90257 vs TME11 0.15 2.67 2.22
I90257 vs TME12 0.11 0.04 8.44**
I90257 vs TME14 0.31 0.24 6.18*
I90257 vs TME4 0.36 0.67 4.53*
I90257 vs TME5 0.34 4.93* 2.47
I90257 vs TM E6 0.07 1.76 1.69
I90257 vs TME7 0.64 0.94 2.69
I90257 vs TME8 0.01 0.32 2.27
I90257 vs TME9 0.18 0.95 6.78**
TME1 vs TME11 0.06 0.98 0.11
TME1 vs TM E12 0.04 0.64 0.31
TME1 vs  TM E14 0.00 0.18 0.10
TME1 vs TME4 0.00 0.00 0.03
TME1 vs TME5 0.02 3.11 0.02
TME1 vs TME6 0.15 0.45 0.09
TME1 vs TME7 0.01 0.03 0.00
TME1 vs TME8 0.62 0.20 0.05
TME1 vs  TME9 0.07 0.05 0.09
* P-value <0.05
** P-value <0.01
1 2 0
T a b le  4 .1 4  C on tin ued
Female parent
Resistant male 58308 130001 I30555 I30572
TM E4 vs TME5 0.01 2.91 0.44
TM E4 vs TME6 0.14 0.43 0.78
TM E4 vs TME7 0.03 0.03 0.10
TM E4 vs TME8 0.58 0.20 0.70
TM E4 vs TME9 0.08 0.04 0.07
TM E5 vs TME6 0.12 1.55 0.06
TME5 vs TME7 0.09 2.88 0.06
TME5 VS TME8 0.58 5.70* 0.01
TME5 vs TME9 0.05 2.01 1.11
TME6 vs TME7 0.36 0.30 0.20
TME6 vs  TME8 0.17 1.56 0.02
TME6 vs  TME9 0.03 0.15 1.60
TME7 vs TME8 1.00 0.47 0.12
TME7 vs TME9 0.30 0.00 0.30
TME8 vs TME9 0.33 0.44 1.72
TME11 vs TME12 0.00 3.26 4 .29*
TME11 vs TME14 0.05 1.88 2.07
TM E 11 vs TME4 0.05 0.92 1.19
TME11 vs  TME5 0.03 0.79 0.09
TME11 vs TME6 0.02 0.13 0.00
TME11 vs TME7 0.17 0.76 0.27
TME11 vs TME8 0.28 2.48 0.03
TME11 vs TME9 0.00 0.47 2.72
TME12 vs TME14 0.04 0.12 0.36
TME12 vs TME4 0.04 0.62 0.68
TME12 vs TME5 0.02 6.23* 2.21
TME12 vs TME6 0.02 2.30 2.76
TME12 vs TME7 0.13 1.04 0.93
TME12 vs TME8 0.21 0.19 2.98
TME12 vs TME9 0.00 0.91 0.65
TME14 vs TME4 0.00 0.17 0.06
TME14 vs TME5 0.02 4.31* 0.90
TME14 vs TME6 0.13 1.15 1.36
TME14 vs TME7 0.01 0.37 0.29
TME14 vs TME8 0.52 0.00 1.31
TME14 vs TME9 0.06 0.36 0.00
* P-value <0.05
** P-value <0.01
121
Table 4.15 Chi-square values for differences in the mean distribution of disease severity 
scores of F-i progenies in crosses involving resistant and susceptible male parents when 
crossed with each of the female parents 130001 resistant to CMD, I30555 susceptible to 
CMD and I30572 resistant to CMD
Male res is tan t source 58308
Female parent 
130001 I30555 I30572
130001 vs I30555 0.99 0.08 6.37* 1.10
I30555 vs 58308 0.59 0.46 5.22* 0.02
130555 vs I30572 1.54 2.51 10.99** 1.68
14(2)1425 vs TME10 0.00 0.00 1.30
14(2)1425 vs TME117 0.14 0.16 8.83**
14(2)1425 vs TME2 1.10 1.62 6.63*
14(2)1425 vs  TME31 0.01 0.04 5.98*
14(2)1425 vs TME41 0.04 0.15 2.01
160142 vs  TM E10 2.33 1.51 0.01
160142 vs TM E117 0.83 2.14 1.80
160142 vs TME2 7.63** 8.84** 1.48
160142 vs TME31 1.34 1.93 0.99
160142 vs TME41 1.28 0.51 0.03
I90257 vs TME10 0.74 2.18 2.70
I90257 vs TME117 0.18 3.07 11.46**
I90257 vs TME2 3.20 8.93** 9.05**
I90257 vs TME31 0.55 2.77 8.23**
I90257 vs TME41 0.32 1.08 3.44
TME1 vs TME10 0.01 0.71 0.02
TME1 vs TME117 0.07 1.50 0.63
TME1 vs TME2 0.98 7.92** 0.58
TME1 vs TME31 0.02 1.16 0.34
TME1 vs TME41 0.01 0.08 0.00
TME2 vs TME31 0.77 2.42 0.13
TME2 vs TME4 1.96 7.86** 1.19
TME2 vs TME5 2.81 1.01 2.77
TME2 vs TME6 3.03 6.34* 3.32
TME2 vs TME7 1.92 8.74** 1.46
TME2 vs TME8 4.33* 13.22** 3.44
TME2 vs TME9 4.00 5.88* 1.20
TME4 vs TME41 0.00 0.07 0.07
TME10 vs TME11 0.21 0.03 0.12
TME10 vs TME12 0.17 2.87 2.22
TME10 vs TME14 0.01 1.55 0.89
TME10 vs TME5 0.13 1.15 0.00
TME10 vs TME6 0.43 0.04 0.08
TME10 vs TME7 0.00 0.56 0.05
TME10 vs TME8 1.14 2.09 0.02
TME10 vs TME9 0.36 0.31 1.07
TME11 vs TM E 117 0.00 0.24 7.50**
TME11 vs TME2 2.26 4.15* 4.75*
TME11 vs TME31 0.16 0.04 4.23*
TME11 vs TME41 0.03 0.59 0.62
TME12 vs TME2 1.77 13.00** 0.21
TM E12 vs TME31 0.12 3.17 0.01
TME12 vs TME41 0.02 1.23 1.17
TME14 vs TME31 0.02 1.97 0.44
TME14 vs TME41 0.01 0.48 0.27
* P-value <0.05
** P-value <0.01
122
T a b le  4 .1 5  C on tin ued
M ale res is tan t source 58308
Female
130001
parent
I30555 130572
TME31 vs TME4 0.05 1.05 0.75
TME31 vs TME5 0.09 0.35 2.24
TME31 vs TME6 0.30 0.26 2.76
TME31 vs TME7 0.01 0.91 0.99
TME31 vs TME8 0.86 2.55 2.97
TME31 vs TME9 0.20 0.62 0.73
TME41 vs TME5 0.00 2.55 0.17
TME41 vs TME6 0.11 0.19 0.40
TME41 vs TME7 0.07 0.01 0.01
TME41 vs TME8 0.53 0.62 0.31
TME41 vs TM E9 0.05 0.00 0.35
TME117 vs TME12 0.00 3.17 0.34
TM E 117 vs TME14 0.06 2.18 1.43
TME117 vs TME4 0.06 1.29 1.95
TME117 vs TME5 0.04 0.03 4 .15*
TME117 vs  TME6 0.03 0.52 4 .72*
TME117 vs TME7 0.23 1.16 1.92
TME117 vs TME8 0.33 2.73 5.46*
TME117 vs TME9 0.00 0.88 2.62
* P-value <0.05 
** P-value <0.01
123
Table 4. 16 Chi-square values for differences in the mean distribution of disease 
severity scores of Ft progenies in crosses involving susceptible male parents when 
crossed with each of the female parents 130001 resistant to CMD, I30555 susceptible to 
CMD and I30572 resistant to CMD
Male parent 130001
Female
I30555 I30572
TME2 vs TME31 0.77 2.42 0.13
TME2 vs TME41 2.47 8.09** 1.64
TM E10 vs TME2 1.71 5.38* 2.84
TME10 vs TME31 0.00 0.13 2.26
TM E10 vs TME41 0.11 0.40 0.15
TME31 vs  TME41 0.08 0.74 1.22
TME10 vs TME117 0.28 0.36 4.19*
TME117 vs TME2 3.01 0.88 0.00
TM E117 vs TME31 0.19 0.08 0.19
TME117 vs TME41 0.04 1.01 2.77
* P-value <0.05 
** P-value <0.01
124
The minimum number of effective factors segregating for resistance to CMD in 
crosses between the various sources of resistance and susceptible parents is 
presented in Tables 4.17 to 4.19. The number of factors responsible for resistance 
varied from cross to cross and from environment to environment. Where little (less 
than 0.2) or no genetic variance was detected, the number of effective factors was 
not estimated.
Among the crosses involving resistant parents (Table 4.17), the number of effective 
factors ranged from 2 to 3 across environment, 2 to 4 in Mokwa 1998, 2 to 6 in 
Ibadan 1998 and from 2 to 3 in Ibadan 1999. In most cases, both parents contributed 
effective factors to the cross. Generally both resistant parents contributed equal 
numbers of effective factors except for a few cases were the better parent 
contributed more.
Effective factors for crosses between resistant and susceptible parents ranged from 2 
to 4 across environments, 2 to 7 in Mokwa 1998, 2 to 6 in Ibadan 1998, and 2 to 7 in 
Ibadan 1999 (Table 4.18). Generally, both the resistant and susceptible parents 
contributed effective factors. In the majority of cases, the resistant parent contributed 
more factors to resistance. However, in six crosses with TMS 130001 as the resistant 
parent and seven crosses with TMS I30555 as the susceptible parent, both parents 
contributed equally in at least one environment.
In crosses between susceptible parents, the effective factors ranged from 2 to 4 
across environments, 2 to 5 in Ibadan 1998 and 2 to 4 in Ibadan 1998 (Table 4.19). 
Generally, both parents contributed equally to resistance, although in some cases, 
TMS 130555 the better parent, contributed more.
4.2.3.4. Number of effective factors
125
Table 4.17 Number of effective factors (NE), number of effective factors contributed by better parent (NBP) and number of effective factors
contributed by poorer parent (NPP) in F, crosses between susceptible parents
C ross
Across environments Mokwa 1998 Ibadan 1998 Ibadan 1999
NE NBP NPP NE NBP NPP NE NBP NPP NE NBP NPP
I30555 x I30555 2 1 1 _ _ _ - - . 2 1 1
I30555 x TME2 2 1 1 2 1 1 3 2 1 - - -
I30555 x  TME10 2 1 1 - - - 3 2 1 2 1 1
I30555 x TME31 2 1 1 - - - 2 1 1 2 1 1
I30555 x TME41 4 2 2 - - - 5 3 2 3 2 1
I30555 x TM E117 3 2 1 - - - 2 1 1 4 3 1
- not estimated due to little or no genetic variance
126
Table 4.18 Number of effective factors (NE), number of effective factors contributed by better parent (NBP) and number of effective factors
contributed by poorer parent (NPP) in Fi crosses between resistant parents
Across environm ents M o kw a 1998 Ibadan 1998 Ibadan 1999
Cross NE NEBP NEPP NE NEBP NEPP NE NEBP NEPP NE NEBP NEPP
130001 x 130001 2 1 1 - - 2 1 1 2 1 1
130001 x I30572 2 1 1 - - - - - 2 1 1
130001 x I58308 2 1 1 - - - - 2 1 1
130572 x  130001 3 2 1 - - 6 4 2 2 1 1
I30572 x I30572 2 - - - - - 2 1 1
I30572 x I58308 2 1 1 - - - - - 2 1 1
I58308 x  130001 2 1 1 - - - - - 2 1 1
I58308 x  I30572 2 1 1 - - - - 2 1 1
I58308 x 158308 2 1 1 - - - 2 1 1
130001 x 14(2)1425 2 1 1 2 1 1 2 1 1 3 2 1
130001 X 160142 2 1 1 2 1 1 2 1 1 2 1 1
130001 x  I90257 2 1 1 2 1 1 2 1 1 3 2 1
130001 x  TME1 2 1 1 2 2 1 1 2 1 1
130001 x  TME4 2 1 1 2 1 1 2 1 1 2 1 1
130001 x  TME5 2 1 1 4 2 1 1 2 1 1
130001 x TME6 2 1 1 2 1 1 2 1 1 2 1 1
130001 x TME7 2 1 1 2 1 1 2 1 1 2 1 1
130001 x TME8 2 1 1 2 1 1 2 1 1 2 1 1
130001 x TME9 2 1 1 2 1 1 4 1 1 2 1 1
130001 x TM E11 2 1 1 2 1 1 4 1 1 2 1 1
130001 x TME12 2 1 1 2 1 1 2 1 1 2 1 1
130001 x TME14 2 1 1 2 1 1 2 1 1 2 1 1
- not estimated due to little or no genetic variance
127
Tab le  4 .1 8  C on tin ued
Cross
Across environments Mokwa 1998 Ibadan 1998 Ibadan 1999
NE NEBP NEPP NE NEBP NEPP NE NEBP NEPP NE NEBP NEPP
I30572 X 14(2)1425 2 1 1 3 2 1 1
130572 x 160142 2 1 1 - - 2 1 1 2 1 1
I30572 x  I90257 3 1 - - - 4 2 1 1
I30572 x TME1 2 1 1 2 1 1 2 1 1 2 1 1
I30572 x  TME4 2 1 1 2 1 1 3 2 1 2 1 1
130572 x  TME5 2 1 1 2 1 1 2 1 1 2 1 1
130572 x TME6 2 1 1 2 1 1 2 1 1 2 1 1
130572 x  TME7 2 1 1 2 1 1 3 2 1 2 1 1
130572 x  TME8 2 1 1 4 2 2 2 1 1 2 1 1
130572 x  TME9 2 1 1 4 2 2 2 1 1 2 1 1
I30572 xTM E11 2 1 1 3 2 1 - 2 1 1
I30572 x  TME12 2 1 1 3 1 1 2 1 1 2 1 1
I30572 x TME14 2 1 1 2 1 1 2 1 1 2 1 1
- not estimated due to little or no genetic variance
128
Table 4.19 Number of effective factors (NE), number of effective factors contributed by better parent (NBP) and number of effective factors
contributed by poorer parent (NPP) in F, crosses between resistant and susceptible parents
Cross
Across environments M o kw a 1998 Ibadan 1998 Ibadan 1999
NE NEBP NEPP NE NEBP NEPP NE NEBP NEPP NE NEBP NEPP
130001 x I30555 3 2 1 - - - - 3 2 1
I30572 X I30555 2 1 1 - - - 2 1 1
I58308 x 130555 2 1 1 - - - - - 2 1 1
130001 x TME2 2 1 1 2 1 1 2 1 1 -
130001 x TME10 2 1 1 - - - 3 2 1 2 1 1
130001 x TME31 2 1 1 - - 2 1 1 2 1 1
130001 x  TME41 2 1 1 - - - 2 1 1 2 1 1
130001 XTM E117 2 1 1 - - - 2 1 1 2 1 1
130572 x  TME2 3 2 1 3 2 1 4 2 2 3 2 1
130572 xT M E 10 4 3 1 - - - 6 4 2 3 2 1
I30572 x TME31 4 2 1 3 3 1 2 1 1 5 3 2
I30572 x  TME41 2 1 1 - - - 2 1 1 2 1 1
I30572 x  TME117 5 4 1 - - 3 2 1 7 6 1
130555 x 130001 2 1 1 - - - 3 1 2 2 1 1
I30555 x  I30572 2 1 1 - - - - - 2 1 1
I30555 x  I58308 2 1 1 - - - - - 2 1 1
I30555 x 14(2)1425 4 2 7 4 3 - - - 2 1 1
130555x160142 2 1 1 - - - 2 1 1 2 1 1
I30555 x I90257 3 2 1 - - - 3 2 2 3 2 1
- not estimated due to little or no genetic variance
129
Tab le  4 .19  C on tin ued
Cross
Across environments Mokwa 1998 Ibadan 1998 Ibadan 1999
NE NEBP NEPP NE NEBP NEPP NE NEBP NEPP NE NEBP NEPP
I30555 x TME1 3 2 1 _ 5 3 2 2 1 1
I30555 x TME4 2 1 1 - - - 2 1 1 2 1 1
I30555 X TME5 2 1 1 3 2 1 2 1 1 2 1 1
I30555 x TME6 3 2 1 - - - 2 1 1 3 1
I30555 X TME7 4 3 1 4 3 2 5 3 2 2 1 1
I30555 X TME8 3 2 1 - - - 5 3 2 2 1 1
I30555 x TME9 2 1 1 2 1 1 2 1 1 2 1 1
I30555 xTM E11 3 2 1 3 2 1 5 2 2 3 1
I30555 x TME12 3 2 1 - - - 5 3 2 2 1 1
I30555 x TME14 3 2 1 5 3 2 2 1 1 2 1 1
not estimated due to little or no genetic variance
130
4.3 Determination of DNA markers associated with resistance to CMD
4.3.1 Resistance screening and selection of plants for bulk segregant analysis
The mean disease severity scores for shoot tips, whole plant, leaf severity, and the 
presence or absence of ACMV viruses of the F1 plants in the mapping population is 
presented in Table 4.20. Based on these results, plants were classified as highly 
resistant (HR), resistant (R), susceptible (S), and highly susceptible (HS). A total of 
ten plants from the highly resistant and highly susceptible classes respectively were 
selected for BSA. The plants that made up the resistant bulk were: 9,11, 22, 29, 36, 
40, 43, 45, 54, and 59, while those that made up the susceptible bulk were: 1, 8, 28, 
31,33, 42, 56, 62, 63 and 67.
4.3.2 Genomic DNA quality and concentration
The DNA extraction method yielded genomic DNA of good quality and quantity. 
There was no smearing when checked by agarose gel (0.8%) electrophoresis. On 
average, 254 ng of DNA per 100 mg of fresh leaves was obtained based on the 
estimates given with the flourometer. The ratio of OD26onm to OD28onm determined by 
spectrophotometry ranged from 1.86 to 2.00 for the samples.
4.3.3 Cassava mosaic virus strains in mapping population
The ACMV specific primer 1 (ACMV-F1 /ACMV-R1) and 2 (ACMV-ALF/ACMV- 
ARO/R.) detected ACMV in 61 and 37 samples, respectively (Table 4.20). The 
ACMV specific primer 1 detected the virus in 29 resistant samples (24 HR and 5 R 
samples) while ACMV specific primer 2 detected the virus in only nine resistant 
samples (8 HR and 1 R). Thus ACMV 1 detected the virus in 20 more samples. 
ACMV specific primer 2, on the other hand, detected the virus in one resistant 
sample, sample 48, which was not detected by ACMV specific primer 1.
131
Table 4.20 Reaction of F-i plants in mapping population to CMD severity in their shoot 
tips (MST), whole plant (MWP), mean severity of first 10 leaves plant (MLS), total mean 
response presence or absence of ACMV virus and their classification (Class)
Plant num ber MST MWP MLS Mean ACM V ACM V 
1 2
Status
1 4 4.5 3.8 4.10 1 1 HS
2 2 4 2.9 2.97 1 1 S
3 1.5 1.5 1.3 1.43 0 0 HR
4 1 2 1.1 1.37 1 1 HR
5 3.5 4 3.8 3.77 1 1 HS
6 1 1 1 1.00 1 0 HR
7 1 2 1 1.33 1 0 HR
8 3.5 4.5 3.4 3.80 1 1 HS
9 1 1 1 1.00 1 0 HR
10 1 1 1 1.00 1 1 HR
11 1 1 1.1 1.03 1 0 HR
12 1.5 2 1 1.50 1 0 R
13 2.5 3.5 1.2 2.40 1 0 R
14 3 4 3.9 3.63 1 1 HS
15 4 4 3.4 3.80 1 1 HS
16 4 4 3 3.67 1 1 HS
17 1 1 1 1.00 0 0 HR
18 1 2.5 1.3 1.60 1 0 R
19 4 4 3.9 3.97 1 1 HS
20 3.5 4 3.7 3.73 1 1 HS
21 1 1.5 1.2 1.23 1 1 HR
22 1 1 1 1.00 1 1 HR
23 3 4 4 3.67 1 1 HS
24 3.5 4 3.2 3.57 1 1 HS
25 3 3.5 3.2 3.23 1 0 S
26 3.5 4 3 3.50 1 1 HS
27 4 4 3.2 3.73 1 0 HS
28 4.5 4 3.7 4.07 1 1 HS
29 1 1 1 1.00 1 0 HR
30 1.5 1.5 1 1.33 1 0 HR
31 4.5 4.5 4.3 4.43 1 1 HS
32 2 4 2.4 2.80 1 1 S
33 4.5 4 3.5 4.00 1 0 HS
34 3.5 4 3.4 3.63 0 0 HS
S, susceptible HS, highly susceptible
R, resistant HR, highly resistant
ACMV1 ACM V specific primer (ACMV-F1/ACMV-R1)
ACMV2 ACMV specific primer (ACMV-AL/F/ACMV-ARO/R)
1 =Presence of v irus product
0=Absence o f virus product
132
T a b le  4 .2 0  con tin u ed
Plant num ber MST MWP MLS Mean ACM V ACM V 
1 2
Status
35 1 1 1 1.00 1 1 HR
36 1 1 1 1.00 0 0 HR
37 1 1.5 1.1 1.20 0 0 HR
38 3.5 4 3.5 3.67 1 1 HS
39 1.5 2 1.3 1.60 1 1 R
40 1 1 1 1.00 1 0 HR
41 2.5 4 2.2 2.90 1 1 S
42 4.5 4 3.6 4.03 1 1 HS
43 1 1 1 1.00 0 0 HR
44 1 1.5 1 1.17 1 0 HR
45 1 1 1 1.00 1 0 HR
46 1 1.5 1 1.17 1 0 HR
47 1 1 1 1.00 1 0 HR
48 1 1 1 1.00 0 1 HR
49 1 1 1.1 1.03 1 0 HR
50 1 1.5 1.2 1.23 1 1 HR
51 2 2 1 1.67 1 0 R
52 1 1 1 1.00 0 0 HR
53 1.5 3 1.2 1.90 1 1 S
54 1 1 1 1.00 1 0 HR
55 3.5 3.5 1.2 2.73 1 1 S
56 4 4 3.3 3.77 1 1 HS
57 3.5 4 3.8 3.77 1 1 HS
58 1 1 1.1 1.03 1 0 HR
59 1 1 1 1.00 1 0 HR
60 2 3 2.5 2.50 1 1 S
61 3.5 4 3.4 3.63 1 0 HS
62 4.5 4.5 4.3 4.43 1 1 HS
63 4 4 4 4.00 1 1 HS
64 3.5 4 3.3 3.60 1 1 HS
65 1 1 1 1.00 1 0 HR
66 3.5 3.5 2.7 3.23 1 1 S
67 4 4.5 3.7 4.07 1 1 HS
68 1 1 1 1.00 1 1 HR
69 1 1 1.1 1.03 1 0 HR
S, susceptible HS, highly susceptible
R, resistant HR, highly resistant
ACMV1 ACM V specific primer (ACMV-F1/ACMV-R1)
ACMV2 ACMV specific primer (ACMV-AL/F/ACMV-ARO/R)
1 = Presence o f virus product
0= Absence o f virus product
13 3
The ACMV primer ACMV 1, also detected the virus in 32 susceptible samples (24 HS 
and 8 S) and ACMV 2 detected the virus in 28 susceptible samples (21 HS and 7 S). 
The virus was not detected in six resistant samples, 3, 17, 36, 37, 43 and 52 and in 
one susceptible sample 34 by either of the two ACMV specific primers. Plate 4.1 
shows the amplification products of some of the samples showing the presence of 
ACMV with the two primers. The EACMV specific primer and the EACMV/UgV primer 
did not detect the presence of the virus in any of the samples.
4.4 Screening parents and bulks for polymorphic markers
4.4.1 Random amplified fragment length polymorphism analysis
A total of 403 markers were recorded from 107 of 142 RAPD primers which were 
used to screen the parents and bulks. The number of bands per primer ranged from 
1-8, averaging 4 markers per primer. Forty-eight of the 107 primers were polymorphic 
between the two parents and three of these exhibited polymorphism between the two 
bulks. The three primers, however, did not distinctively distinguish between the 
resistant and susceptible DNAs when used to screen the individual member of the 
two bulks. An example of the amplification products using RAPD primers, K11, K14, 
K15, K16, and M05, on the parents and bulks is shown in plate 4.2. In this example, 
all the primers except primer K15 were polymorphic between the two parents and 
none of the primers were polymorphic between the bulks.
134
Panel 1 
Panel 2
1 2 3 4 5 6 7 8 91011 12 13 14 15 16 171819 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45
Plate 4.1 Amplification products of section of mapping population showing the presence of ACMV with two 
ACMV specific primers
Panel 1: ACMV specific primer ACMV-F1/ACMV-R1F1/R1 
Panel 2: ACMV specific primer ACMV-AL/F/ACMV-ARO/R 
Lane 1:1 Kb marker
Lane 2 to 45: Example of F1 progenies with or without ACMV PCR products
1 3 5
1 2  3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
P1 P2 R S P1 P2 R S P1 P2 R S P1 P2 R S P1 P2 R S
Plate 4.2 Amplification profile generated by some RAPD primers on, P1 (TMS I30555), 
P2 (TME7), R (resistant bulk) and S (susceptible bulk).
Lanes 1-4, Operon Primer K11;
Lanes 5-8 Operon Primer K14;
Lanes 9-12, Operon Primer K15;
Lanes 13-16, Operon Primer K16;
Lanes 17-20 Operon Primer M05.
136
4.4.2 Amplified fragment length polymorphism analysis
The average number of bands and the percentage polymorphism between the 
parents and the bulks for the different AFLP systems is shown in Table 21.
With the EcoR1//Wse1 analysis, amplification products were obtained from 57 primer 
pairs of the 64 used to screen the two parents (TMS I30555 and TME7) all of which 
were polymorphic between the two parents. The number of bands detected after the 
silver staining ranged from 23 to 108, giving an average of 68.5 bands per primer 
pair. A total of 3,906 bands were generated, out of which 843 were polymorphic 
averaging 14.8 polymorphic bands per primer pair, and an overall of 21.58% 
polymorphism. For the bulked DNAs, amplification products were obtained from 54 of 
the EcoRMMse'\ +3 primer combinations. These yielded a total of 2,860 bands 
averaging 51.07 per primer pair, of which only 119 bands were polymorphic, 
averaging 2.13 polymorphic bands per primer pair and giving 4.16% level of 
polymorphism between the bulks.
The Apa1+C/7ag1+A templates screened with the 40 combinations of /\pa1+3 and 
7ag1+3 primers yielded a total of 1,903 bands from 35 primer pairs. The total number 
of bands ranged from 19 to 93, averaging 54 bands per primer. There were 401 
bands, which were polymorphic between the parents, giving an average of 7.36 
polymorphic bands per primer pair and 21.13% polymorphism on the whole for this 
system. A total of 314 of the bands were polymorphic between the two bulks, giving 
16.5% polymorphism among the bulks. However, when the were used to screen the 
members of the two bulks, the results were very erratic, and none of the polymorphic 
primer pairs showed a clear distinction between the two groups.
The /4pa1+A/Taq1+C templates screened with the 6 primer combinations (/\pa 1+ATT 
and Taq1+ CAC, CAG, CAT, CTA, CTG or CTT) yielded a total of 475 bands ranging 
from 62 to 88 bands per primer pair, with an average of 79 bands per primer pair. A
137
total of 73 of the 475 bands were polymorphic between the parents, giving 15.4% 
polymorphism between the parents and 8 were polymorphic between the bulks, 
giving 1.68% polymorphism between the bulks.
138
Table 4.21 Average number of bands, level of polymorphism among parents (TMS 
130555, TME7) and bulks with different AFLP systems
Preamplification Template Average 
number o f bands
Level o f Po lym orph ism  (%) 
Parents Bulks
E coM +A /M se+C 68.5 21.58 4.16
/Apa1+C/Taq1+A 54.4 21.13 16.50
Apa 1+A/Taq1+C 79.2 15.4 1.68
/4pa1+C/Taq1+C 45.3 17.81 1.43
Psf1 +A/T aq1+C 51 9.87 0.3
1 3 9
The /\pa1+C/Taq1+C templates screened with 12 primer combinations (involving 
Apa  ^ + CTA or CTG and Taq1 + CAC, CAG, CAT, CTA, CTG or CTT) yielded a total 
of 494 bands for the parents, averaging 45 bands per primer pair. The number of 
polymorphic bands between the parents ranged from 3 to 18, averaging 8.8 
polymorphic bands per primer pair, and the total number of polymorphic bands was 
88 giving 17.81% polymorphism. For the bulks, a total of 491 bands were scored and 
8 were polymorphic between the bulks, giving 1.43% polymorphism.
The Pst 1+ A/Taq 1+C templates screened with 24 of the primer combinations 
involving Psf1+ ACA or ACC and Taq1 + CAC, CAG, CAT, CTA, CTG or CTT yielded 
a total of 1,165 bands ranging from 20 to 83 bands and an average of 51 bands per 
primer. A total of 115 of the total number of bands were polymorphic between the two 
parents, giving 9.87% polymorphism between the parents, while three of the bands 
were polymorphic between the two bulks, giving 0.3% polymorphism.
4.4.3 Simple Sequence Repeats
The number of bands (microsatelliite alleles) generated from the SSR markers 
ranged from 1 to 3 per sample. Fifteen markers were polymorphic for both the 
parents and the bulks. One SSR primer SSR30, which had three alleles (a, b, and c) 
among the two parents, consistently distinguished between the two DNA bulks and 
the members of each bulk group. The polymorphic marker (allele “a”), designated as 
SSR30-180, was present in the resistant landrace TME7 but absent in the 
susceptible improved line I30555. The band was also present in all the members of 
the resistant bulk, but absent in the members of the susceptible group (Plate 4.3).
140
1 2 3 4 5 6 7 8 9 10 11 12 131415 16 17 18 19 20 21 22 23 24 25
Plate 4.3 Amplification products of SSR primer 30 in parents, bulks and bulk members. 
Lane 1, TME7;
Lane 2, resistant bulk;
Lanes 3 to 12 samples 9, 11, 22, 29, 36, 40, 43, 45, 54 and 59;
Lane 13, TMS I30555;
Lane 14, susceptible bulk,
Lane 15 to 24 samples 1, 8,14,19, 23, 42, 57, 60, 62 and 63;
Lane 25, 50bp marker.
Long arrow showing allele a, marker SSR30-180 
Shorter arrows showing two other alleles b and c
141
4.5 Association of DNA markers with CMD resistance in Fi mapping 
population
4.5.1 Genetic Linkage analysis
The nine EcoRMMse'\+3 primers, three Psf1/Taq1+3 primers and 15 SSR primers, 
which were used in genotyping the Fi mapping population, gave a total of 138 
polymorphic bands for linkage analysis. The resistant male parent TME7 contributed 
73 and the susceptible female parent TMS I30555 contributed 65 markers of the 138 
loci generated from the AFLP and SSR analysis.
Plates 4.4 and 4.5 show examples of sections of the Fi mapping population following 
selective amplification with the primer E-AAC/M-CAT and P-ACC/T-CTA. An example 
of the amplification products of three SSR primers, SSR6, SSR70 and SSR7, on a 
section of the F-i mapping population is also shown in plate 4.6.
Prior to linkage analysis, the markers were subjected to chi-square analysis to test 
their conformity to the 1:1 and 3:1 inheritance ratios. The chi square analysis 
revealed that at the 5 and 1 % probability levels, 38 (58.46%) of the markers 
contributed by TMS I30555 and 44 (60.27%) contributed by the resistant parent 
TME7 were segregating for the1:1 segregation ratio of single dose markers. The 
results further showed that 10 (15.38%) of the markers donated by the female and 
four markers (5.48%) donated by the male parents were significant for 3:1 
segregation ratio of double dose markers (Tables 4.22 and 4.23).
142
Plate 4.4 A cross section of AFLP products of parents TMS 130555 and TME7 and progenies following
amplification with primers E-ACC/M-CAT. Arrows showing polymorphic bands
Lane 1: TMS I30555
Lane 2: TME7
Lane 3 to 17: F.,s 1 to 15
Plate 4.5. Section of AFLP products for primer P-ACC/T-CTA. Arrows showing polymorphic 
bands
Lane 1: 50bp marker;
Lane 2: TMS I30555;
Lane 3; TME7;
Lanes 4 to 38; F1 plants 36 to 69
144
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Panel III
Plate 4.6. Example of multiplexing PCR products of three SSR primers 
Panel I: SSR6, Panel II: SSR70, Panel III: SSR7 
Lane 1: TMS I30555 
Lane 2: TME7
Lanes 3 to 2: F1 population, samples 1 to 17 
Arows showing polymorphic bands
145
Table 4.22 Segregation analysis for conformity to 1:1 and 3:1 segregation ratio of
polymorphic DNA markers scored in mapping population contributed by P1 I30555
Alle le Chi va lue fo r 1:1 ratio Chi va lue fo r 3:1 
ratio
SSR132-A 0.00 68.00
E -ACC /M -CAT-175 0.01 73.06
SSR 124-2 0.01 65.06
SSR6/290 0.06 76.24
E-ACC /M -CAT-130 0.13 57.52
E-ACA /M -CAT-130 0.13 79.54
E-ACC /M -CAT-500 0.36 90.45
SSR56-B 0.37 48.49
E-ACA /M -CAT-145 0.47 51.88
E-ACC /M -CTC-175 0.53 46.12
P-ACC /T-CT A -210 0.53 46.12
E -ACC /M -CAT-364 0.56 90.25
P-ACTIT -CTT  -440 0.71 99.84
P-ACT/T-CTT-435 0.71 43.84
E -ACT/M -CAG -245 0.73 41.93
E -ACT/M -CAG -240 0.73 41.93
SSR157-B 0.75 40.02
E-ACT/M -CAG -210 0.94 103.76
E-ACC/M -CTC-650 0.94 103.76
SSR70/320 0.94 39.76
E-ACA/M -CAC-340 1.17 109.70
E-ACA/M -CAT-360 1.40 12.60
E-ACC /M -CT G -450 1.52 32.06
E-ACC/M -CAT-775 2.31 80.26
E-AAG /M -CT G -220 2.45 130.80
E-ACA/M -CAC-650 2.45 26.80
E-ACT/M -CAG -190 2.52 25.09
P-ACC/T-CAG-380 2.52 25.09
E-AAG/M -CTG-560 2.88 135.53
E-AAG /M -CTG -182 3.26 142.04
P-ACT/T-CTT-520 3.26 142.04
E-AAG/M-CTC-170 3.76 19.06
E-ACA/M -CAC-545 3.76 147.06
SSR7/220 3.76 19.06
E-ACA/M -CAC-280 4.19 153.75
E-ACC/M -CT G-48 4.91 13.64
E-AAG /M -CTC -195 6.39 178.57
E-ACA/M -CAT-480 6.45 7.81
Tabulated chi va lue at 5% probability 3.84 
Tabulated chi va lue at 1% probability 6.63
146
T a b le  4.22 C on t
Alle le Chi va lue fo r 1:1 ratio Chi va lue for 3:1 
ratio
E -ACT/M -CAG -208 7.12 184.47
E -ACC /M -CAT-370 7.67 7.67
E -ACA /M -CAT-305 7.81 5.23
E -ACA /M -CAT-160 7.90 6.58
E-AC A /M -C  AT -240 9.94 3.76
E -ACC /M -CTC -210 9.94 3.76
SSR119-B 9.94 211.76
E -ACC /M -CT  G -350 10.24 2.97
E-ACA /M -CAT-330 10.90 1.61
E -AAG /M -CT G -181 11.53 226.12
E-AAG /M -CT G -620 12.19 233.75
E -ACC /M -CAT-745 12.19 233.75
E -ACC /M -CAT-650 12.60 1.40
E -AAG /M -CTC -350 13.24 0.94
E -AAG /M -CTG -110 13.93 248.71
E-ACC /M -CAT-220 15.78 0.13
E-ACC /M -CTC -180 17.75 280.01
SSR41-A 17.75 280.01
P-ACCT/-CAG-340 18.28 280.13
E-ACA/M -CAT-155 18.69 221.77
P-ACC/T-CAG-175 20.43 0.73
P-ACC/T-CAG-240 23.53 322.12
E -ACC /M -CTC-215 25.94 339.76
P-ACC/T-CAG-170 39.76 19.06
SSR50-A 47.52 32.06
E -ACC /M -CAT-180 53.93 40.71
E-ACC /M -CT G -260 54.55 44.18
Tabulated chi va lue a t 5% probability  is 3.84 
Tabulated chi va lue a t 1% probability  is 6.63
147
Table 4.23 Segregation analysis for conformity to 1:1 segregation ratio of polymorphic
DNA markers scored in mapping population contributed by P2 TME7
Alle le Chi va lue fo r 1:1 ratio Chi va lue for 3:1 
ratio
E -ACC /M -CTG -180 0.00 66.00
E-AAG /M -CTG -550 0.01 65.06
E-ACC /M -CAT-725 0.01 65.06
E-ACC /M -CTC-450 0.01 65.06
E-ACT /M -CAG -17 5 0.01 71.06
SSR132-B 0.01 63.06
E-ACT/M -CAG -250 0.06 60.24
SSR41-B 0.13 57.52
E-AAG /M -CTG-850 0.24 52.94
E-ACA /M -CAT-150 0.24 84.94
P-ACC /T -CTA-190 0.24 84.94
SSR56-A 0.24 84.94
E-AAG /M -CTC -55 0.36 50.45
SSR115-A 0.36 50.45
P -ACC /T-CTA-145 0.37 88.49
SSR119-C 0.37 88.49
E-ACA /M -CAT-170 0.53 94.12
SSR6/340 0.53 46.12
SSR124-A 0.53 46.12
E-ACC/M -CTC-240 0.71 43.84
P-ACC/T-CTA-470 0.73 97.93
E-ACA/M -CAT-94 0.94 103.76
SSR30-180 1.17 37.70
E-AAG /M -CTG -180 1.47 113.88
E-ACA /M -CAT-118 1.47 33.88
E-ACC/M -CAT-95 1.75 32.01
SSR7/210 1.75 120.01
E-ACC/M -CAT-105 2.88 23.53
E -ACA /M -CAT-175 2.88 23.53
E-ACT/M -CAG-230 2.88 135.53
E-ACT/M -CAG-92 2.88 23.53
E-ACC/M -CAT-795 2.88 23.53
E-ACC/M -CTG-45 2.97 21.88
SSR119-A 3.36 140.43
P-ACT/T-CTT-175 3.76 19.06
E-ACT/M -CAG-95 4.31 16.25
E-ACC/M -CTC-225 4.76 159.06
P -ACT/T-CTT-110 5.23 13.93
E-ACC/M -CAT-570 5.88 171.53
E-ACA/M -CAC-290 5.88 11.53
SSR50-B 6.06 10.24
E -ACC/M -CAT-83 6.39 10.57
E -AAG /M -CTC -470 6.39 10.57
SSR50-C 6.58 177.33
Tabula ted chi va lue at 5% probability is 3.84 
Tabulated chi va lue at 1% probability is 6.63
148
T a b le  4 .2 3  co n tin u ed
Alle le Chi va lue fo r 1:1 ratio Chi va lue fo r 3:1 
ratio
SSR157-A 7.33 183.33
E-ACC /M -CTC -120 8.47 5.88
E-ACC /M -CTG -55 8.73 196.91
E-ACC /M -CTG -325 9.33 204.31
E-ACC /M -CTG -635 10.88 218.52
E-ACA /M -CAT-345 10.94 152.76
E -ACA /M -CAT-230 12.55 233.21
E-ACC /M -CAT-352 13.24 0.94
P-ACC /T-CTA-350 13.24 240.94
E-AAG /M -CT C-275 13.93 248.71
E -ACC /M -CTG -315 15.52 256.06
P-ACC /T-CTA-280 17.00 272.00
P-ACC /T-CAG -425 17.00 272.00
P-ACT /T -CTT-75 17.00 0.00
E-ACA/M -CAT-77 21.00 21.00
E -ACC /M -CTG-52 21.88 1.52
E-ACC /M -CTG-230 22.70 313.81
P-ACT n  -CTT-66 24.36 330.45
E-ACC/M -CTG-570 25.09 331.36
E-ACC/M -CTG-540 25.09 331.36
E-ACC /M -CTG -105 26.73 340.91
E-ACC/M -CTG-215 29.33 359.33
E-ACC /M -CTG -150 32.06 378.24
E-ACC/M -CTG-680 36.76 415.06
P-ACC/T-CAG-220 36.76 415.06
P-ACT/T-CTT-55 40.71 443.84
E-ACC/M -CTG-370 45.15 467.60
P-ACC/T-CAG-440 49.47 497.88
E-ACC/M -CTG-748 68.00 612.00
Tabulated chi va lue a t 5% probability  is 3.84 
Tabulated chi va lue a t 1% probability  is 6.63
14 9
Initially, all 138 markers were subjected to linkage analysis. This generated seven 
linkage groups with about 80% of the markers in linkage group-l. The 96 markers 
segregating for the two genetic ratios were then tested with different linkage analysis 
strategies (i.e. single female and male maps, combined map, single maps with 
inverted loci, and combined maps with inverted loci). The most suitable results were 
obtained for the combined maps. With all 96 markers, only 36 markers mapped in 15 
linkage groups and 12 groups had just two markers.
Since the backcross configuration of single-dose segregating markers is desirable in 
mapping heterozygous F-, populations (Liu, 1998); the 14 markers segregating for the 
3:1 ratio were eliminated and finally, 82 of the 138 markers (59.42%) and the 
converted CMD resistance locus CMDRL, were used to construct the linkage map.
The loci mapped in 14 linkage groups with six groups having only two markers each 
(Table 4.24). The 14 linkage groups ranged in size from 16.8 to 227.6 cM and the 
number of markers per linkage group correlated with the map distances per linkage 
group (^=0.98). The distance between adjacent loci in centimorgans on the map 
ranged from 14.9 to 73.2 cM with an average of 33.66 cM, even though the distance 
threshold set with the LOD of 3 and linkage criteria of 0.33 was 39.6 cM. On the 
whole, the genetic linkage map spanned 1,111.4 cM of the genome. The AFLP 
primer pair E-ACT/M-CAG contributed six markers in four linkage groups (I, V, VI and 
VII), which was the highest number of markers contributed by any of the AFLP primer 
pairs.
The SSR30-180 marker and an AFLP marker, E-ACC/M-CTC-225, flanked the 
converted CMDRL on linkage group-X (Fig. 4.2). Both markers were donated by the 
resistant parent TME7. The map distances that separated CMDRL from SSR30-180 
and E-ACC/M-CTC-225 were 14.9 and 31.10 cM respectively.
150
Distance in centimorgan Marker Name
E-ACC/M-CAT225
31.10 cM
CMDRL
14.9 cM
SSR130/180
Fig. 4.2 Linkage group-X of the linkage map showing the location of the resistance to 
CMD, defined as the CMDRL
152
4.5.2 QTL analysis
Simple linear correlation revealed significant association between six markers, 
including CMDRL and the CMD response traits (Table 4.25). The marker SSR30-180 
had the highest linear association after CMDL for all the response traits and their 
mean effect. Two markers SSR119-A and SSR119-C, which mapped on linkage 
group-XIII and were donated by the resistant male parent, had the next largest 
association with CMDRL and the CMDR traits. SSR30-180 and E-ACC/M-CTC-225 
had positive correlation with CMDRL and negative correlation with the traits while, 
SSR119-A and SSR119-C had negative correlation with CMDRL and positive 
correlation with the responses. An unassigned marker, E-AAG/M-CTG-550, had a 
weak correlation (p<0.05) with mean whole plant severity.
Significant marker-trait association was detected by regression analysis for the four 
markers SSR30-180, E-ACC/M-CTC-225, SSR119-A and SSR119-C and the CMD 
resistance responses (Table 4.26). The r2 values, which indicated the proportion of 
the phenotypic variation explained by the markers, were highest for SSR30-180 in all 
cases. It accounted for 57.41% of the phenotypic variation in the population for the 
mean response, followed by SSR119-C which accounted for 37.83%, SSR119-A 
accounted for 32.24% then E-ACC/M-CTC-225 accounted for 22.5% of the total 
phenotypic variation in the population for resistance to CMD.
153
Table 4.25 Simple linear correlation coefficient between markers and CMD responses
Marker Linkage Group CMDL MST MWP MLS Mean
E-ACC/M-CTC-225 X 0.453*** -0.413*** -0.481*** -0.531*** -0.486***
SSR30180 X 0.723“ * -0.738*** -0.769*** -0.715*** -0.762***
SSR119-A XIII -0.548*** 0.537*** 0.588*** 0.556*** 0.577***
SSR119-C XIII -0.619*** 0.610*** 0.591*** 0.617*** 0.622***
CMDL X 1.000**** -0.898*** -0.943*** -0.910*** -0.942***
E-AAG/M-CTG-550 UM -0.228 -0.208 -0.275* -0.241 -0.241
* p< 0.05 probability level 
*** p< 0.005 probability level 
UM: unassigned marker 
MST: Mean shoot tips,
MWP: Mean whole plant,
MLS: Mean severity of first 10 leaves plant 
Mean: Total mean response
154
Table 4.26 Association between markers SSR30-180 and E-ACC/M-CTC-225 with all 
CMD resistance traits (All traits), mean shoot-tip severity (MST), mean whole plant 
severity (MWP), and mean leaf severity (MLS) in mapping population
Linkage G roup CMD response r(%) F-ratio p-value
E-ACC /M -CTC -225 X MST 15.79 13.75 0.0004
MWP 22.01 20.19 0.0001
MLS 27.13 26.32 0.0001
Mean 22.5 20.74 0.0001
SSR30180 X MST 53.76 80.07 <0.0001
MWP 58.48 96.76 <0.0001
MLS 50.43 70.19 <0.0001
Mean 57.41 92.65 <0.0001
SSR119-A XIII MST 27.82 27.21 <0.0001
MWP 33.59 35.4 <0.0001
MLS 29.87 29.96 <0.0001
Mean 32.24 33.36 <0.0001
SSR119-C XIII MST 36.36 39.78 <0.0001
MWP 33.99 36.02 <0.0001
MLS 37.18 41.25 <0.0001
Mean 37.83 42.38 <0.0001
E-AAG/M -CTG-550 UM MST 3.77 3.66 ns
MWP 2.88 3.01 ns
MLS 6.18 5.48 ns
Mean 4.42 4.15 ns
UM= unassigned marker 
Ns= not s ign ificant
155
4.6 Association of DNA marker SSR30-180 with CMD resistance in other 
cassava clones
There were differences between some resistant clones and TME7 with respect to the 
marker SSR30-180. The SSR primer SSR30 detected four different alleles (a, b, c, 
and d) among all the 23 cassava clones used as parents and checks in the genetic 
NCD II and diallel experiments as opposed to the three it detected between the two 
parents TMS I30555 and TME7. The resistant clones 58308, TMS I30572, and the 
landrace TME8, did not have the marker SSR180-30 (allele “a”). The primer amplified 
a different allele (allele “d”) for these clones, suggesting that different alleles may be 
involved in cassava resistance to CMD (Plate 4.7). The susceptible clones TME31 
and TME117 had alleles similar to those of TME7.
156
1 2 3 4 5 6 7 8 9 10 11 12 1314 1516 17 18 192021 2223
Plate 4.7 Amplification product of SSR 30 on 23 cassava clones
Lane 1: TME1 
Lane 5: TME6 
Lane 9: TME10 
Lane 13: TME31 
Lane 17: TMS I30555 
Lane 21: TMS 160142
Lane 2: TME2 
Lane 6: TME7 
Lane 10: TME11 
Lane 14: TME41 
Lane 18: TMS I30572; 
Lane 22: TMS I90257
Lane 3: TME4 
Lane 7: TME8 
Lane 11: TME12 
Lane 15: TME117 
Lane 19: 58308 
Lane 23: 91/02324
With alleles a, b and c as seen in plate 4.3 and new allele d occurring in some clones.
Lane 4: TME5 
Lane 8: TME9 
Lane 12: TME14 
Lane 16: TMS 130001 
Lane 20: TMS 14(2)1425
157
Chapter Five 
Discussion
Resistance to CMD from the resistant genetic stock, clone 58308 which has been the 
main source of resistance to CMD in breeding is polygenic, recessive and inherited 
largely in an additive manner (Hahn and Howland, 1972). Polygenic inheritance has 
also been reported for some of the African landraces, which are potential new 
sources of resistance to the disease (Lokko et al., 1998). However, a major dominant 
gene responsible for resistance in an African landrace has also recently been 
reported (IITA, 2000; Akano et al., 2002). In this study, further progress was made in 
understanding the genetics of resistance to CMD from the 58308 source and among 
the landraces.
Genotype by environment interaction on inheritance of resistance
Fargette et al. (1994) had previously shown that CMD responses are environment 
dependent. The significant GXE detected for most of the genotypic effects in the 
NCD II experiment and GCA effects in the diallel indicate a lack of stability of the 
genetic effects measure across environments. The differences in ranks of the 
AUDPC estimates across environments further demonstrated the lack of stability of 
the CMD symptom severity across environments. The importance of GXE interaction 
detected in this study shows that none of the genotypes can be effectively assessed 
in any single environment. Therefore, multiple environmental evaluation, which has 
been recommended for traits with significant environmental interaction effect 
affecting them, is required to obtain precise genetic information (Zhang and Kang,
1997).
158
Relative importance of GCA/SCA
The relative importance of GCA or additive gene effect in predicting progeny 
performance was estimated from the ratios of the mean square components 
associated with variance of GCA and SCA. The closer this ratio is to 1, the greater 
the chances of predicting progeny performance based on GCA (Baker, 1978; Kang, 
1994). From the results, GCA was relatively more important in determining progeny 
performance in the NCD II experiment and the GCA effect due to both the females 
and the males was important, although this was environment-dependent. The 
predominance of GCA in the NCD II was further reflected by the significant 
correlation between parental means and GCA, which implies that progeny 
performance can be predicted from the parental performance. The significance of the 
female by male interaction (SCA), however, suggests the presence of some non­
additive components, which was detected in some crosses.
Results from previous diallel experiments indicated that resistance to CMD derived 
from clone 58308 was inherited largely in an additive manner (IITA, 1973; Hahn et 
al., 1980; 1989; Msabaha, 1983). In the diallel experiment, additive gene effect which 
was due mainly to clone TMS I30572 and 58308, was important in the Ibadan 1998 
environment. However, the magnitude of the ratios of the mean square components 
associated with variance of GCA and SCA, across environments, in Mokwa 1998 and 
Ibadan 1999 indicate a predominance of SCA which implies that non-additive genetic 
effect was more important in determining progeny performance. A predominance of 
SCA is expected among crosses involving inbred lines (Matzinger and Kempthorne, 
1956). Although cassava is a heterozygous species, three of the parents used in the 
diallel experiment are known to have common ancestry and, therefore, the 
assumption of no inbreeding did not hold. The population could therefore be likened
159
to a diallel cross among inbred lines and explains the predominance of SCA in the 
diallel experiment.
In this study, negative GCA effects of a parent indicated that the Fi crosses involving 
that parent are more resistant than the average F-i crosses. Negative values of 
clones indicated a larger contribution towards resistance whiles positive significant 
value suggest a contribution towards susceptibility. The best general combiner in the 
NCD II, which contributed most to resistance, was the improved clone TMS I90572. 
The clone TMS I30572 was the best general combiner among the females and the 
landraces TME8 and TME9 were good general combiners in the Ibadan 1999 
environment. TMS I30572 together with TMS 130001 were the best general 
combiners among the parents in the diallel experiment. TMS 130001, however, was a 
poor general combiner in the NCD II experiment. In both experiments TMS I30555 
was a poor general combiner. The resistant genetic stock, clone 58308, was also a 
poor general combiner in the diallel and also had a significant but positive maternal 
effect. However, in earlier studies, 58308 exhibited good parental value for resistance 
to CMD (IITA, 1973). In another study conducted in three environments, the 
landraces TME8, TME9, and TME11 were the best general combiners for resistance 
to CMD while TMS 130001, TMS I30555, and TMS I30572 were poor general 
combiners (Lokko et al., 1998).
The results of the present study compared with those from previous studies (IITA, 
1973; Lokko et al., 1998) show that the GCA of a clone is dependent on the genetic 
constitution of the population within which it is being tested as well as the 
environment. They further show that the resistant parents identified as potential 
sources of resistance for breeding need to be selected carefully after progeny testing 
to determine the most suitable parents.
160
The overall best general combiner detected in the NCD II experiment, TMS 190257 
should be used in breeding programmes to enhance resistance to CMD. Other 
clones such as TMS 130572, TMS 160142 TMS 14(2)1425, the landraces TME8 and 
TME9 could also be used in breeding programmes to enhance resistance but the 
progenies would have to be tested in different environments to select resistant 
phenotypes. The high heritability estimates obtained in the two experiments shows 
that generally, significant progress in breeding for resistance to CMD can be made by 
selection. Ogbe, (2001) recommended TMS 130001 and TME8 as sources of 
resistance for breeding, based on their ability to express two types of resistance to 
CMD, resistance to virus movement and resistance to virus multiplication. However, 
based on the results of this study, TMS 130001 is a poor general combiner for 
resistance to CMD. It should, therefore, be used in combination with the resistant 
clones, which have been shown to have good combining ability.
Significant and negative SCA effects were desirable for resistance. A cross with 
significant and negative SCA implied that this cross was more resistant than the 
average GCA of the parent clones and a cross with significant and positive SCA 
implied that this cross was more susceptible than the average GCA of the parent 
clones. The three crosses with negative and significant SCA effects across 
environments were more resistant than expected based on the average performance 
of the parents and in Ibadan 1999, two crosses were more resistant than the average 
performance of the parents. Although two good general combiners were involved in 
one of these crosses, the parents in the other crosses were not the best general 
combiners. This implies that there is no general relationship between the GCA of a 
parent and the SCA of its cross. Furthermore, out of the six crosses that were more 
susceptible than the average of their parents across environments, only three had 
one of their parents as the poor general combiner (TME2). In Ibadan 1999, none of 
the parents involved in the three crosses with significant and positive SCA, were poor
161
general combiners. This shows that progeny testing is also required to select the best 
parental combinations for CMD resistance.
The results of the diallel also revealed that there is no direct relationship between the 
level of resistance in a parent and the SCA effect of its cross. TMS 130555 X TMS 
130572 involving a good general combiner with negative and significant GCA (TMS 
130572) and a poor general combiner (TMS 130555) with positive and significant GCA 
were more resistant in Mokwa 1998 than the average performance of the parents. In 
Ibadan 1999, two crosses TMS 130001 X 58308 and TMS I30555 X TMS I30572 
were more resistant than the average performance of the parents while their 
reciprocals were not, which further suggests that progeny testing is required to 
determine resistant genotypes in a cross. The selfs of TMS I30555 and 58308, on the 
other hand, were more susceptible than the average performance of the parents in 
Mokwa 1998. Although it is expected that selfing in cassava reduces hybrid vigor, 
significant and positive SCA was the general trend in all the selfs and for the selfs 
TMS I30555 X TMS 130555 and 58308 X 58308 in all environments.
The significant reciprocal and maternal effects detected in the Ibadan 1998 
environment and maternal effects detected in Mokwa 1998 environment were due to 
the significant and positive maternal effects of clone 58308. Maternal effect refers to 
the effect of the maternal genotype or tissues on a trait of its offspring, while non- 
maternal effect represent the effect on a trait of non-nuciear or extranuclear genetic 
factors or the interaction between products of nuclear genes and plasmagenes (Kang 
et al., 1999). Maternal effect due to 58308 was also environment-dependent, 
therefore progenies with 58308 as the mother would have to be tested in different 
environments to select for higher levels of resistance to CMD.
162
In this study, heritability was estimated from the mean square components of the 
crosses. These values represent estimates of narrow sense heritability since the 
genetic variances among half sibs equals to covariances of the half sibs, which 
represent primarily additive genetic variance contained in the phenotypic variance 
among the half sib families (Nguyen and Sleper, 1983). The results show that 
significant progress in breeding for resistance to CMD can be made from selection. 
These estimates of 51.6% and 56.6% in the NCD II and diallel respectively, are 
comparable to those reported in previous studies, which varied from 30% to 68% 
(Hahn et al. 1980a). The difference in the magnitudes of the heritability estimates 
indicates that the composition of the genotypes, and environment in which they are 
evaluated were also important.
Heterosis
The significance of the parents’ mean squares in the two experiments is an indication 
of the diverse variability among the parents used in the study. Significant differences 
among the male parents and resistant parents in the NCD II experiment suggests 
that the African landraces could serve as potential sources of exotic genes for 
resistance. The significant and negative mid-parent and high parent heterosis 
estimated for CMD severity in some crosses is further evidence that the landraces 
could serve as good sources of resistance.
Heterosis occurs when there is allelic interaction or overdominance at some or all 
alleles, when the parents differ in gene frequency and when directional dominance 
exists (Kang, 1994; Falconer and Mackay, 1997). However, heterosis would vary 
from cross to cross. The significance of the contrast parent versus cross in the 
individual environments for both experiments suggests the presence of average mid-
Heritability o f CMD
163
parent heterosis. Heterosis was also environment dependent, therefore crosses need 
to be evaluated in several environments to determine the most heterotic cross.
Non-addivtive gene effect due to additive by additive epistatsis has been proposed 
as the cause of heterosis in self pollinating species (Gravois, 1994;) whiles in 
outcrossers both dominance effect and additive by additive epistatsis contribute to 
SCA causing heterosis (Holland, 2001). The means for the various crosses however, 
showed various intermediate values, which suggested polygenic inheritance and the 
absence of dominance. No dominance is assumed if the mean in a first generation 
cross is equal to the mid-parent (Hallauer and Miranda, 1988). These suggest that 
heterosis detected in some of the crosses may be due to allelic interaction (Xiao et 
al., 1995; Falconer and Mackay, 1997; Yu et al., 1998 ).
Genetic differences based on parental responses to CMD
The PCA 1 versus PCA 2 scatter-plots illustrated the genetic relationship among the 
clones, with genotypes within a cluster group being genetically related.
The results implied that the four resistant landraces TME7, TME8, TME9 and TME12 
are genetically similar to TMS 130001. The tiers in the ranks of AUDPC values for 
either shoot tip or whole plant in at least one environment for these clones further 
suggest genetic similarities among them. Genetic similarities between TMS I90257 
and the landraces TME1 and TME6 and between TMS I30572 and TMS 160142 and 
landraces TME11 were also established. These genetic similarities suggest that the 
clones in the same genetic group may exhibit the same level of resistance 
mechanism. As was expected, the susceptible clones were genetically distinct from 
the resistant clones.
Mode of inheritance of resistance to CMD
In many cases, resistance to plant viruses is under a simple genetic control involving 
a single dominant or recessive gene (Fraser, 1990). However, there are reports in
164
the literature indicating that resistance to some plant viruses is under complex 
genetic control (Hahn and Howland, 1972; McMullen et al., 1994; Caranta and 
Plaloix, 1996; Melchinger et al., 1998). This study also confirmed the polygenic mode 
of resistance to CMD.
The resistant parents used in this study did not represent the genotypic extremes but 
had varying levels of resistance in the different environments. Similarly, the 
susceptible parents exhibited varying levels of susceptibility. The segregating Fi 
crosses between the various sources of resistance and susceptible parents and 
between the resistant and susceptible parents also exhibited varying levels of 
resistance and susceptibility to CMD, suggesting multiple factors with genes from 
both parents. The presence of resistant phenotypes in the crosses involving 
susceptible by susceptible parents imply that resistance to CMD in the clones studied 
is recessive. These results are similar to those of Hahn and Howland (1972) who 
reported polygenic and recessive inheritance for resistance to CMD. In crosses 
involving two resistant parents, the presence of susceptible progenies also suggest 
that the genes for resistance in these parents are non-allelic. Wang et al. (1998) 
tested allelism forresistance to the soybean mosaic virus and showed that if 
resistance genes from the two parents are non-allelic and not linked, a proportion of 
the progenies from their crosses would be susceptible.
Transgressive segregation
Transgressive segregation is due either to mutants induced by hybridity, the 
unmasking of recessive genes in heterozygote parents, or complementary action of 
genes with additive effects (Rick and Smith, 1953; Xu et al., 1998). Since mutation 
induced by hybridity rarely happens (Xu et al, 1998), the likely explanation for the 
transgressive segregants (TS) detected in this study would be the unmasking of 
recessive genes in heterozygote parents or complementary action of genes which 
were detected in the study. The presence of transgressive segregation in this study is
165
further evidence for the polygenic control of resistance to CMD among the cassava 
clones studied that involved different alleles in different parents and is cumulative for 
degree of resistance.
The lack of TS for CMD severity in the diallel population suggests that similar genetic 
mechanisms affect the cassava clones used as parents in the diallel (Fenster and 
Ritland, 1994). Another explanation for the lack of TS in the diallel is that there is 
linkage among loci. Linkage could also result in a reduced frequency of transgressive 
segregates for a trait (Lawrence and Frey, 1976).
Gene complementarity
The main purpose of testing for gene complementarity is to examine relationships 
among genes from different sources of a trait in order to determine whether the 
genes are allelic. Gene complementarity among sources of resistance to disease and 
pest have been reported in oats, soybean and sorghum (Dixon et al., 1991; Fox et 
al., 1997; Wang etal., 1998).
Significant differences in the mean distribution of F-i progeny disease severity scores 
of the 15 sources of resistance to CMD, when these were crossed with resistant 
clones TMS 130001, TMS I30572, and the susceptible clone TMS I30555, are an 
indication of allelic differences in their genes for resistance. As was expected, there 
were no significant differences between the resistant improved clones, which were 
derived from clone 58308. However, the improved clones 14(2)1425, 160142 and 
I90572 were different from three, one and five resistant landraces respectively. While 
no significant differences were detected due to reciprocal effects, the significant 
differences detected when either TMS I30555 or TMS I30572 was the female parent 
indicates that the expression of resistance is influenced by the nature of the female 
parent. The resistant landrace TME5 for instance, showed significant differences 
from three improved clones, TMS 14(2)1425, TMS 160142 and TMS I90572 and
166
landraces TME8, TME12 and TME14 when these were crossed with TMS I30555. 
This indicates the presence of minor allelic genes, which are influenced by specific 
crosses.
Significant differences between the resistant and susceptible clones imply that 
although the susceptible parents contributed some effective factors to resistance, 
their allelic effects were different from those of the resistant parents. Interestingly, the 
test did not reveal significant differences between susceptible clones TME2 and 
TME10, which were also different from TME117.
Allelic differences imply the possibility of TS and the presence of heterosis. Due to 
the polygenic and recessive nature of resistance to CMD reported earlier, Hahn et al. 
(1979) recommended population improvement and recurrent selection in breeding, 
as this would accumulate desirable gene combinations that would be difficult for the 
pathogens to circumvent in the long term. The use of non-allelic clones to induce TS 
would further increase resistance in a breeding population.
Number of effective factors
In this study, the parent clones did not represent the genotypic extremes. 
Furthermore, the analysis of variance revealed the presence of non-additive 
components, thus, the use of the range between the extreme Ft progenies in the 
estimation of effective factors.
The minimum number of effective factors affecting a quantitative trait is expected to 
range from one to two times the haploid number of the species. However, linkage, 
gene interaction such as epistasis and the presence of transgressive segregants 
could cause a reduction in the estimated number of effective factors (Fenster and 
Ritland, 1994). Studies have shown that the absence of linkage assumed in the 
Castle-Wright formula is not usually attained (Zeng et al. 1992; Lynch and Walsh, 
1998) and would account for an underestimation of the number of effective factors.
167
The few effective factors estimated for some crosses could be attributed to any of 
these factors since transgressive segregants and the presence of gene interaction 
from the significance of the SCA effects were detected. The number of effective 
factors was not estimated in crosses where there was little or no genetic variance, 
which may be due to the absence of differences in the genes or differences in gene 
action of the parents.
The number of effective factors estimated for the resistant by resistant crosses 
suggests that both the resistant improved clones and resistant landraces have equal 
potential in contributing resistant genes to their progenies. Furthermore, the presence 
of plus factors observed in progenies of resistant by resistant crosses indicates that 
the sources of resistance carry different genes that complement each other. The 
contribution of effective factors by both resistant and susceptible parents also 
indicates that resistance to CMD can be enhanced by continually combining elite 
genotypes regardless of their resistance status.
There are no previous reports on the number of effective factors responsible for 
resistance to CMD. Hahn et al. (1980b), inferred from the polygenic nature of 
resistance to CMD that resistance must be attributed to the combined action of the 
number of loci which are linked on a chromosome or a set of chromosomes of a 
genome or genomes. They, however, suggested that since cassava is genetically 
heterozygous and probably an allotetraploid, a study on the genetic mechanism of 
resistance to CMD would be complicated and difficult.
Virus strain causing CMD symptoms
The results of this study showed that ACMV was responsible for the disease 
symptom in the mapping population. Although mixed infections of ACMV and 
EACMV have been reported in West Africa (Ogbe et al., 1999; Offei et al., 1999), 
none was detected in the study using two primers. The ACMV primer ACMV-
168
F1/ACMV-R1 was more efficient in detecting the virus in the samples in that it 
detected the virus in more samples than the primer ACMV-AL F/ACMV-ARO R. 
However, the two primers did not detect the virus in a susceptible plant, which maybe 
due to a high concentration of DNA template, which would inhibit the amplification 
reaction.
The presence of the virus in some of the resistant samples suggests that field 
resistance was not necessarily an indication of resistance to virus infection. The A 
genome of the gemini-viruses encode a protein required for their replication and must 
recruit the remaining DNA replication mechanism from the host plant, while the B 
genome is responsible for spread and symptom production (Estessami et al., 1991; 
Fontes et al., 1992). Since DNA replication is part of the natural growth and 
development, it is possible that the virus is able to replicate and probably even 
spread in the resistant plant but the subsequent disease symptoms are inhibited.
Using field evaluation, axial bud inoculation and PCR, Ogbe (2001) concluded that 
field resistance as shown by lack of symptoms was not necessarily an indication of 
resistance to virus infection, but could be partly due to lack of virus multiplication, 
which suggests that field selection of resistance should be complemented with PCR 
or inoculation test.
Development of molecular markers associated with CMD resistance
The detection of multiple genes for virus resistance using segregation analysis alone 
is not efficient because of the differences due to genotype by environment interaction 
(McMullen and Louie, 1989). To determine molecular markers associated with 
resistance to CMD, three marker systems, RAPDs AFLPs and SSRs were employed, 
using bulk segregant analysis (BSA) and linkage analysis.
169
RAPD markers have the potential of detecting unique regions in the genome due to 
the random amplification of sites within the genome. In BSA, where, DNA samples 
are pooled from individuals which are similar for a particular trait, RAPD markers are 
particularly useful since it is expected that the low frequency allele(s) in the pooled 
DNA sample will not be amplified (Liu, 1999; Michelmore et al., 1991). Using BSA, a 
RAPD marker was found to be linked with high cyanogenic potential (CNP) in 
cassava (Dossou-Yovo, 2000). In this study, however, none of RAPD markers 
assayed were able to detect any significant differences between the resistant and 
susceptible samples.
The AFLP experiments, as expected, generated several data points per assay. The 
different AFLP experiments were performed with BSA to generate large number of 
markers and the possibility of some association with resistance to CMD. Reamon- 
Buttner et al. (1998), for instance, detected nine AFLP markers tightly linked to the 
sex locus of the dioecious plant, Asparagus officinalis L, using the silver staining 
technique. In this study however, BSA did not reveal tight linkages between any 
AFLP marker and resistance to CMD.
Wong et al. (1999) reported that cassava has a GC rich genome, therefore, 
restriction enzymes such as Taq1, Apa1 and Pst1, which recognise GC sites would 
be more suitable in AFLP analysis of cassava. The highest number of bands was 
detected with one of the Apa1 experiments and the highest number of polymorphic 
bands was also obtained with another Apa 11 Taq 1 experiment. The number of bands 
detected in this study with the silver staining detection technique which is affected by 
quality of reagents and water purity, were good compared with other AFLP studies on 
cassava which used radioactive detection (Fregene et al., 2000; Roa et al., 1997).
Wong et al. (1999) reported high levels of polymorphism among cultivars using the 
ApaHTaql restriction enzymes. Although the Apa1ITaq1 experiment, generated the
170
highest number of polymorphism in the bulks, they gave very erratic results, which 
were not repeatable, thus no useful markers were detected for the subsequent 
genotyping of the entire mapping population. This could be due to the high degree of 
non-specific amplification due to the differences in the +3 selective primers and the 
selective amplification template, which can cause mismatches. Certain mismatches 
at the terminal 3’ would allow amplification without a reduction in the amplification 
yield (Kwok et al., 1990).
Mapping molecular markers associated with CMD resistance
The linkage analysis established associations between resistance to CMD and two 
markers, SSR30-180 and E-ACC/M-CAT225, and suggests the presence of a locus 
for resistance to CMD between the two markers. Both markers were donated by the 
resistant parent TME7; that shows that their linkage with CMD resistance is in 
coupling phase. The distance of 14.9 cM between SSR30180 and the converted 
resistance locus, however, suggests that the detected marker is not tightly linked to 
resistance. Its high association with resistance in the mapping population may be due 
to inheritance of that segment from TME7 in the resistant progeny. This indicates the 
need to saturate this region of the genome with more DNA markers, to obtain a 
marker that is more tightly linked to resistance to CMD.
While BSA and linkage analysis confirmed the presence of a locus associated with 
resistance to CMD on linkage group-X, the marker trait association test further 
showed the presence of another region on linkage group X-lll associated with CMD 
resistance. The correlation coefficients for SSR30-180 and E-ACC/M-CTC-225 with 
CMD responses were negative while those for SSR119-A and SSR119-C were 
positive. This implies that the CMD resistance genes associated with these two 
regions of the genome are different.
171
The SSR marker SSR30-180 which explained about 58% of the total variation has 
also been identified in another cassava population from an African landrace, TME3 
(known in Western Nigeria as 2nd Agric). In that study, the marker mapped on linkage 
group R of the male cassava framework map with a recombination distance of 8 cM 
from a dominant cassava resistance gene CMD2 (IITA, 2000; Akano et al., 2002). 
The differences in the two-recombination values may be due to the smaller 
population size in this study. The results of the two studies, however, imply that the 
resistant gene CMD2 is related to the genes or a gene for resistance in TME7 and 
the other resistant clones with the marker. However, the presence of the marker in 
two susceptible clones also suggests that resistance genes in the landraces may be 
genetically related to non-resistance genes, with few differences in their sequences.
The differences between some resistant clones and TME7 with respect to the marker 
SSR/30-180 further suggest that different alleles may be involved in resistance to 
CMD. These differences could be utilised in breeding to enhance resistance, as was 
shown from the significant and negative SCA effects, significant and negative 
heterosis, and positive TS in some crosses.
Currently, further development of SSR markers involving untranslated regions of 
cassava ESTs for SSR repeats are underway (Mba et al., 2001). Another project 
involves fine mapping the region of the cassava framework map at CIAT, followed by 
contig mapping with bacterial artificial chromosome (BAC) libraries (Fregene et al., 
2001). Since the markers, which had the highest association with CMD resistance 
were developed from microsatellite loci with CT repeats (Mba et al., 2001), exploring 
more markers based on such repeats could facilitate efforts at fine mapping and 
isolating CMD resistance genes. Alternatively, high throughput markers such as EST 
obtained from cDNA clones of known resistance genes and SNPs, may be employed 
in mapping CMD resistance genes. Due to the high frequency of SNPs in the
172
genome, there is high possibility of them being closely associated with resistance 
(Lohmann et al., 2000).
Applications of the cassava linkage map
In this study, three strategies for linkage analysis were tested and the single­
combined map gave the most suitable results. In mapping heterozygous species, 
since markers generated in the Fi progenies result from independent meiosis and 
crossing over in the maternal and paternal parents, individual maps are often 
constructed for each parent if the progeny numbers are large enough (Williams,
1998). When this was tested in this study, due to the number of markers for linkage 
analysis, the individual maps generated 5 and 7 linkage groups for the female and 
male respectively which was not adequate. Fregene et al. (1997) suggested that if a 
polyploid origin for the cassava genome is assumed, the expected number of linkage 
groups in cassava would depend on chromosome assortment. Eighteen groups are 
expected where preferential pairing is exclusive, up to 36 where there is mostly 
random pairing, and a number between 18 and 36 where there is a mixture of 
preferential and random pairing (Fregene et al., 1997).
The 14 linkage groups generated for the single-combined map in this study indicate 
that the map is not condensed. Furthermore, the closest distance between any two 
adjacent markers was 14.9 cM. A dense map would have at least one marker in a 5 
cM segment (Liu,1998). This again can be attributed to either insufficient number of 
markers or the size of the mapping population.
Fregene et al. (1997) further described a method to compensate for the random 
assignment of “present or absent" to alternate alleles at a locus and to detect linkage 
in repulsion. In their study, they included inverted scores of the female and male 
mapping data before linkage analysis. This technique also did give a condensed 
map, thus the single-combined map. A key observation in all three techniques
173
(separate male and female maps, single-combined map, and inclusion of inverted 
scores) for linkage analysis, resulted in the markers E-ACC/M-CTC-225 and SSR30- 
180 on the same linkage group and order with CMDL.
Despite the low saturation of the map, the information from it and the map itself can 
be utilised in cassava genetics. The loci were randomly distributed over linkage 
groups since the number of markers per linkage group correlated with the map 
distances per linkage group. Since the mapping population was segregating for 
resistance to CMD and several other traits, the markers could be used to map other 
traits of interest provided quantitative data are available on them. The markers could 
also be applied in studies with similar cassava populations. The AFLP primer pairs E- 
ACT/M-CAG which, contributed markers in four linkage groups could be used in 
experiments such as in diversity studies which require coverage of most of the 
genome. The map could also be used in linkage map pooling with other maps based 
on different parents such as the cassava framework map (Fregene et al., 1997) or 
with one parent in common such as the population studied by Akano et al. (2002) to 
form a composite map of cassava.
A composite map could then be used in planning experiments, to construct a genome 
database, to compare QTL identities in different genetic backgrounds, and for 
comparative mapping with other species. Composite maps have been reported in 
Arabidopsis thaliana and in maize (Beavis and Grant, 1991; Hauge et al., 1993). 
Comparative mapping is a strategy that uses information obtained from the study of a 
species with a simple genome to make inferences about the map position and 
function of genes in species with a complex genome (Brown, 1999). Comparative 
mapping has been used successful in human and mouse genome mapping to 
increase the efficiency of mapping (Liu, 1998). It has been suggested that 
comparative mapping has definite advantages in mapping plant genomes. For 
instance, genes in wheat with a genome size of 17,000 Mb would be mapped from
174
in fo rm a tio n  o b ta in e d  from  a sm a ll m e m b e r o f  th e  G ram in e ae ,  r ice , w ith  a g e n o m e  
s iz e  o f  4 0 0  M b  (G a le  and  D evos , 1998; L iu  1998; B row n  1999).
175
Chapter Six 
Conclusion
The genetic control of resistance to CMD in some African landraces and their 
relationship with the widely used resistant genetic stock, clone 58308, was evaluated 
in two genetic experiments in three environments in Nigeria. This provided 
information on the inheritance of resistance to CMD based on symptom severity in 
the cassava genotypes evaluated and showed how the different sources of 
resistance complement each other.
The importance of GXE interaction in the expression of resistance to CMD in the 
experimental material was demonstrated. The study showed that, depending on the 
population, both additive and non-additive gene effects were involved in the genetics 
of resistance to CMD. In the NCD II experiment involving the landraces, additive 
gene effect was more important while in the diallel involving clone 58308, both 
additive and non-additive gene effects were important. High estimates of heritability 
were obtained in both experiments, which shows that significant progress in breeding 
for resistance to CMD can be made by selection.
The best general combiner was 190257. This clone may be used as parent with any 
other cassava clone to obtain progenies with resistance to CMD. The clones 130572, 
TME8, and TME9, which also exhibited good general combining ability, may be used 
as parent to obtain progenies with resistance to CMD. However, due to the 
importance of GXE, presence of non-additive effects, the trend of heterosis, and the 
transgressive segregants detected, the best parent combinations should be decided 
after progeny testing in multiple environments.
176
The presence of multiple genetic factors or polygenic inheritance for resistance to 
CMD was also demonstrated in the study. The results showed that resistance to 
CMD in the clones studied is also recessive.
The study showed that both the resistant improved clones and resistant landraces 
have potential in contributing resistant factors to their progenies. The presence of 
PTS in the susceptible by resistant crosses and the contribution of effective factors 
by both resistant and susceptible parents indicate that resistance to CMD can be 
enhanced by continually combining elite genotypes. Allelic differences detected 
between some of the cassava clones implies that resistance to CMD in the breeding 
population can be further enhanced by making crosses with the clones with non­
allelic resistant genes, to induce PTS which represent a potential source of novel 
genetic variations.
There is the need to saturate the region of the genome between the marker SSR30- 
180 and CMDL, and between SSR119-A and SSR119-C with more DNA markers, to 
obtain a marker, which is tightly linked with resistance to CMD. High throughput 
marker systems such as EST and SNPs could be applied.
The AFLP primer pair E-ACT/M-CAG, can be used in genetic experiments such as 
diversity studies which require coverage of most of the genome. The linkage map 
could be used to develop a consensus map of cassava that would be of importance 
to cassava research. This would allow researchers to identify informative 
experimental markers, QTL for trait of economic importance and generate a genomic 
database of cassava which could also be used for comparative mapping with other 
species.
1 7 7
References
Adejare, G.O. and R.H.A. Coutts, 1982. The isolation and characterisation of a virus 
from Nigerian cassava plants affected by the cassava mosaic disease and 
attempted transmission of the disease. Phytopathologische Zeitschrift 103:198- 
210.
Adejare, G.O. and R.H.A. Coutts, 1981. Eradication of cassava mosaic disease from 
Nigerian cassava clones by meristem-tip culture. Plant Cell Tissue Organ 
Culture 1:25-32.
Agyare-Tabbi, A., L.F. Pereira and L.R. Erickson, 1997. Isolation and characterisation of 
repetitive and microsatellite DNA sequences in cassava. African Journal of Root 
and Tuber Crops 2:135-137.
Ahiabu, R.K., Y. Lokko, K.E. Danso and G.Y.P. Klu, 1995. Mutagenesis for ACMV 
resistance in a Ghanaian cassava cultivar Bosom nsia. Presented at the fourth. 
F.AO/IAEA/Italy RCM on the improvement of basic food crops in Africa through 
plant breeding including the use of induced mutations. Naples, Italy.
Ahloowalia, B.S., 1995. In vitro mutagenesis for the improvement of vegetatively 
propagated plants. Induced mutations and molecular techniques for crop 
improvement. Proceedings IAEA/FAO International Symposium on the use of 
Induced Mutations and Molecular Techniques for crop Improvement, pg 531- 
541 IAEA, Vienna Austria.
Akano, A.O., A.G.O. Dixon, C.Mba, E. Barrera and M. Fregene, 2002. Genetic mapping 
of a dominant gene conferring resistance to the cassava mosaic disease (CMD). 
Theoretical and Applied Genetics 105:521-525.
Alagianagalingan, M.N. and K. Ramakrishaman, 1970. Studies on virus disease of 
tapioca (M. esculenta Crantz) carbohydrate metabolism. Madgras Agric Journal 
57:55-62.
Allem A.C., 1994. The origin of Manihot Esculenta Crantz (Euphorbiaceae). Genetic 
Resources and Crop Evolution 41:133-150.
Allem A.C. and S.K. Hahn, 1991. Cassava germplasm strategies for Africa. In: Ng. 
N.Q., P. Perrino, F. Attere and H. Zedan (Eds). Crop genetic resources of Africa 
vol. II. Proceedings of an international conference organised by the International 
Institute of Tropical Agriculture (IITA) and the National Research Concil Italy 
(NRC),in association with the International Board for Plant Genetic Resources 
(IBPGR) and the United Nation Environmental Programme (UNEP) held in 
Ibadan Nigeria 17-20 Oct 1998, pg 127-149, Ibadan, Nigeria.
Amma, C. S. E., M.N. Sheela and P.K.T., Pillai, 1995. Combining ability heterosis and 
gene action for three major quality traits in cassava. Journal of Root Crops 
21(1) .24-29
Angel, F, D. Arias, J. Tohme, C. Iglesias and W. Roca 1993. Towards the construction 
of a molecular map of cassava (Manihot esculenta Crantz): comparison of 
restriction enzymes and probe source indetecting RFLPs. Journal of 
Biotechnology 31:103-113.
178
Arunachalam, V., 1976. Evaluation of diallel crosses by graphical and combining ability 
methods. Indian Journal of Genetics and Plant Breeding, 36:358-366.
Asiedu, R., K.V. Bai, R. Terauchi, A.G.O. Dixon and S.K. Hahn, 1992. Status of wide 
crosses in cassava and yam. In: G. Thottappilly, L.M., Monti, D.R., Mohan and 
A.W., Moore (Eds). Biotechnology: Enhancing Research on Tropical Crops in 
Africa. CTA/IITA co-publication, pg. 63-68 IITA, Ibadan, Nigeria..
Awoleye F., M. van Duren, J. Dolezel and F.J. Novak, 1994. Nuclear DNA content in in 
vitro induced somatic polyploidisation cassava (Manihot esculenta Crantz) 
breeding. Euphytica 76:195-202.
Bai, K.B.V., 1982. Cytogenetics of tuber crops. In: IITA (International Institute of Tropical 
Agriculture) Annual report for 1982, Ibadan Nigeria.
Bai, K.B.V., 1987. Recent advances in cassava genetic and cytogenetics. In: Hershey 
C.H. (Ed.) Cassava breeding a multidisciplinary review, Proceedings of a 
workshop held in the Philippines, 4-7 March 1985. CIAT, Cali Colombia.
Bai, K.V., K. Asiedu and A.G.O. Dixon, 1993. Cytogenetics of Manihot species and 
interspecific hybrids. In: Roca, W.M., and A.M. Thro (Eds), Proceedings of the 
first international meeting of the cassava biotechnology network. Cartagena de 
Indias Colombia, 25-28 August, 1992 pg 51-55. Working document No. 123. 
CIAT, Cali, Colombia.
Baird, W.V., R E. Ballard, S. Rajapakse and A.G. Abbott, 1996. Progress in Prunus 
mapping and application of molecular markers to germplasm improvement.
HortScience 31 (7): 1099-1106
Baker, R.J., 1978. Issues in diallel analysis. Crop Science 18(4):533-536.
Barker, B., P. Zambryski, B. Staskawicz and S.P. Dinesh-Kumar, 1997. Signalling in 
plant-microbe interactions. Science 276:726-733.
Barrett, B.A. and K.K. Kidwell, 1998. AFLP-based genetic diversity assesment among 
wheat cultivars from the Pacific Northwest. Crop Science 38:1261-1271.
Beavis W.D and D. Grant, 1991. A linkage map based on information from four F2 
populations in Maize (Zea mays). Theoretical and Applied Genetics 82:636-644.
Beckmann, J.S and M. Soller, 1986. Restriction fragments length polymorphism in plant 
genetic improvement. Oxford Surveys of Plant Molecular and Cell Biology 
3:197-250.
Beeching, J.R., P. Marmey, M.C Gavalda, M. Noirot, H.R. Haysom, M.A. Hughes and 
A. Charrier, 1993. An assessment of genetic diversity within a collection of 
cassava (Manihot esculenta Crantz) germplasm using molecular markers. 
Annals of Botany 72:515-520.
Beil G.M and R.E. Aitkins, 1967. Estimates and specific combining abilities in F, hybrids 
for grain yield and its components in grain sorghum, Sorghum vulgare Pers. 
Crop Science 7:225-228.
1 7 9
Berrie LC  K E. Palmer, E.P. Rybicki and M.E.C. Rey, 1998. Molecular 
’ characterisation of a distinct South African cassava infecting geminivirus. 
Archives of Virology 143:2253-2260.
Best R and G H e n r y ,  1994. Cassava Towards Year 2000. In. International Network for 
' cassava Genetic resources. Colombia, 18-23 August 1992. International Crop. 
Network series No. 10 International Plant Genetic Resources Inst., pg 12-15. 
Rome, Italy.
Bock, K.R., 1994. Control of African cassava mosaic geminivirus by using virus-free 
planting material. Tropical Science 34:102-109.
Bock, K.R. and B.D. Harrison, 1985. African cassava mosaic virus. Association of 
Applied Biologists Description of Plant Viruses No. 297.
Bokanga, M., 1994. Processing cassava leaves for human consumption. Acta 
Horticulturae 375:203-207.
Bokanga, M., 1998. Cassava flour and chips: Commodities for African development in 
the 21st century. Presented at the 7th Triennial symposium of the International 
society for tropical root crops, Africa Branch. 11-17 Oct. 1998, Cotonou, Benin. 
IITA, Ibadan, Nigeria.
Bonierbale, M., C. Guevara, A.G.O. Dixon, N.Q. Ng, R. Asiedu and S.Y.C. Ng, 1997a. 
Cassava. In: D. Fuccillo, L. Sears and P. Stapleton (Eds). Biodiversity in trust 
conservation and use of plant genetic resources in CGIAR centres. Cambridge 
University Press, pg 1-20. Cambridge, UK.
Bonierbale, M.W., R.L. Plaisted and S.D. Tanksley. 1993. A test of populations of alfalfa 
(Medicago sativa) using RAPD molecular the maximum heterozygosity 
hypothesis using molecular markers in tetraploid potatoes. Theoretical and 
Applied Genetics 86:481-491.
Bonierbale, M., A.C. Roa, M.M. Maya, M.C. Duque and J. Tohme, 1997b. Assessment 
of genetic diversity in Manihot species with AFLPs. African Journal of Root and 
Tuber Crops 2:139.
Botstein, D., R.L. White, M. Skolnick, and R.W. Davis, 1980 Construction of a genetic 
map in man using restriction fragment length polymorphism. American Journal 
of Human Genetics 32:314-330.
Brown, T.A. 1999. Genomes. BIOS Scientific Publication, 472pg. New York USA.
Brummer, E.C., 1999. Capturing heterosis in forage crop cultivar development. Crop 
Science 39:943-954.
Bubeck, D.M., M.M. Goodman, W.D. Beavis and D. Grant, 1993. Quantitative trait loci 
controlling resistance to gray leaf spot in maize. Crop Science 33:838-847.
Burr, B. and F.A. Burr, 1991. Recombinant inbreds for molecular mapping in maize: 
theoretical and practical considerations. Trends in Genetics 7:55-60.
Burrow, M.D. and T.K. Blake, 1998.. Molecular tools for the study of complex traits. In: 
A.H. Paterson (Ed) Molecular dissection of complex traits. CRC Press LLC . pq 
13-30. Boca Raton, FL, USA.
180
Byrne, D., 1987. Breeding cassava. In: Russell (Ed), Plant Breeding Reviews 2:73-133.
Caetano-Anolles, D., B.J. Bassam, and P.M. Gresshoff. 1991. DNA amplification 
fingerprinting using very short arbitrary oligonucleotide primers. Biotechnology 9: 
553-557.
Calvalho, I.J.M.B., J.C.M. Cascardo., P.S. Cimera., M.C.M Ribeiro., A.C. Allem and J.J, 
Fialho, 1993. Study of DNA polymorphism in Manihot esculenta and related 
species In: Roca, W.M., and A.M. Thro (Eds), Proceedings of the first 
international meeting of the cassava biotechnology network. Cartagena de 
Indias Colombia, 25-28 August, 1992 pg 56-58. Working document No. 123. 
CIAT, Cali, Colombia..
Calvert, L.A., M.l. Cuervo, M.D. Ospina, C.M. Fauquet, and B.C. Ramirez, 1996. 
Characterisation of cassava common mosaic virus and a defective RNA 
species. Journal of General Virology. 77:525-530.
Caranta, C. and A. Plaloix, 1996. Both common and specific factors are involved in 
polygenic resistance of pepper to several potyviruses. Theoretical and Applied 
Genetic 92:15-20.
Castle, W.E, 1921. An Improved method of estimating the number of genetic factors 
concerned in cases of blending inheritance. Science 54:223.
Changa, 1992. The non-systemicity component of resistance to the African cassava 
mosaic virus in cassava germplasm and its epidemiological implications. MSc 
Thesis University of Ibadan 94 pg.
Chavarriaga-Aguirre, P. M.M. Maya, M.W. Bonierbale, S. Kresovich M.A. Fregene, J. 
Tohme and G. Kochert, 1998. Microsatellites in cassava (Manihot esculenta 
Crantz): discovery, inheritance and variability. Theoretical and Applied Genetics 
97:493-501.
Chavarriaga-Aguirre, P.M. M. Maya, J. Tohme, M.C. Duque, C. Iglesias, M.W. 
Bonierbale, S. Kresovich and G. Kochert, 1999. Using microsatellites, isozymes 
and AFLPs to evaluate genetic diversity and redundancy in the cassava core 
collection and to assess the usefulness of DNA-based markers to maintain 
germ-plasm collections. Molecular Breeding 5:263-273.
Chen, X., S. Temnykn, Y. Xu, Y.G. Cho and S.R. Me Couch, 1997. Development of a 
microsatellite framework map providing genome coverage in rice (Oryza sativa 
L.) Theoretical and Applied Genetics 95:553-567.
Cock, J.H., 1985a. Cassava, new potential for a neglected crop. West View Press, Inc., 
Boulder, Colorado. USA pg 191.
Cock, J.H., 1985b. Cassava: A basic energy source in the tropics. In: Cassava research 
production and utilisation: Preliminary edition. Cassava program, Centro 
International de Agriculture Tropical (CIAT), pg. 1-29. CIAT, Cali, Colombia.
Cock, J.H and R.H. Howeler, 1978. The ability of cassava to grow on poor soils. In: 
Jung G.A. (ed). Crop tolerance to sub-optimal land conditions. American Society 
of Agronomy, pg 145. Madison, Wisconsin, USA.
181
Cockerham, C.C., 1963. Estimation of genetic variances. In: Hanson, W.D. and H.F 
Robinson (ed). Statistical genetics and plant breeding. NRC Publication 982. 
National Academy of Science, National Research Council,, pg 53-93. 
Washington DC, USA.
Cockerham, C.C., 1986. Modifications in estimating the number of genes for a 
quantitative character. Genetics 114:659-664.
Cockerham, C.C. and B.D. Weir, 1977. Quadratic analysis of reciprocal crosses. 
Biometrics 33:187-203.
Collins, N.C., C.A. Webb, S. Seah, J.G. Ellis, S.H. Hulbert and A. Pryor, 1998. The 
isolation and mapping of disease resistance gene analogs in maize. Molecular 
Plant Microbe Interaction 11:968-978.
Comstock, R.E. and H.F. Robinson, 1948. The components of genetic variance in 
populations of biparental progenies and their use in estimating the average 
dominance. Biometrics 4:254-266.
Concibido, V.C., D.A. Lange, R.L. Denny, J.H. Orf and D. Young, 1994. DNA marker 
analysis of loci underlying resistance to soybean cyst nematode (Heterodera 
glycines Ichinohe). Crop Science 34:240-246.
Correa, R.X, M.R. Costa, P.l. Good-God, V.A. Ragagnin, F.G. Faleiro, M.A. Moreira and 
E.G. de Barros, 2000. Sequence characterised amplified regions linked to rust 
resistance genes in common bean. Crop Science 40:804-807.
Cours-Darne, G., 1968. Improving cassava in Africa. The Abidjan Conference. 
Agricultural Research Priorities for Economic Development in Africa. 2: 330-339.
Cox, D.J. and K.L., Frey, 1984. Combining ability and selection of parents for 
interspecific oat matings. Crop Science 24:963-967.
Cukadar-Olemedo, B., J.F., Miller and J.J. Hammond, 1997. Combining ability of the 
stay green and seed moisture content in sunflower. Crop Science 37:378-382.
Danesh, D., S. Aarons, G.E. McGill and N.D. Young, 1994. Genetic dissection of 
oligogeneic resistance to bacterial wilt in tomato. Molecular Plant Microbe 
Interaction 7:464-471.
de Kochko, A.B., R.N, Verdaguer and C.M. Fauquet, 1998. Cassava vein mosaic virus 
(CsVMV), type species for a new genus of plant double stranded DNA viruses? 
Archives of Virology, 143:945-962.
Dixon, A.G.O, P.J. Bramel-Cox and T.L. Harvey, 1990. Diallel analysis of resistance in 
sorghum to greenbug biotype E: antibiosis and tolerance. Crop Science 
30:1055-1059.
Dixon, A.G.O, R. Asiedu and M. Bokanga, 1994. Breeding cassava for low cyanogenic 
potential. Problems progress and prospects. Acta Horticulturae 375:153-161.
Dixon, A.G.O., P.J. Bramel-Cox and T.L. Harvey, 1991. Complementarity of genes for 
resistance to green bug (Schizaphis graminum (Rondani)), biotype E, in 
sorghum ((Sorghum biocolor (L)) Moench). Theoretical and Applied Genetics 
81:105-110.
182
Dorosh, P., 1988. The economics of root and tuber crops in Africa. Resource and Crop 
Management Program (RCMP) Research Monograph No.1 IITA, Ibadan, 
Nigeria.
Dossou-Yovo, I, 2000. Identification de facteurs moleculaires controlant I’accumulation 
de glucosides cyanogeniques dans le manioc: PhD Thesis University of Paris 
IV. 193 pg.
Dudley, J.W., 1993. Molecular markers in plant improvement: Manipulation of genes 
affecting quantitative trait. Crop Science 33:660-668.
Ehrlich, H.A, R. Gibbs and H.H. Kazazian, 1989. Polymerase chain reaction. Cold 
Spring Habour Press. Cold Spring Harbour, NY, USA.
Estessami, P, K. Saunder, J. Watt and J. Stanley, 1991. Mutational analysis of 
complementary sense genes of ACMV DNA-A. Journal of General Virology 
72:1005-1012.
Falconer, D.S. and T.F.C. Mackay, 1997. Introduction to quantitative genetics. 4th 
edition. Longman Scientific and Technical co., Essex, England. 464pg.
FAO, 1997. FAO food outlook. No. 3/4. pg. 21-25.
FAO, 2000. http://www.FAO.org
Fargette, D., L.T. Colon, R. Bouveau and C. Fauquet, 1996. Components of resistance 
of cassava to African cassava mosaic virus. European Journal of Plant 
Pathology 102: 645-654.
Fargette, D., C. Fauquet, E. Grenier and J.M. Thresh, 1990. The spread of African 
cassava mosaic virus into and within cassava fields. Journal of Phytopathology 
130: 289-302.
Fargette, D., C. Fauquet and J.C. Thouvenel, 1988. Yield losses induced by African 
cassava mosaic virus in relation to the mode and the date of infection. Tropical 
Pest Management 34: 89-91.
Fargette, D., C. Fauquet and J.C. Thouvenel, 1985. Field studies on the spread of 
African cassava mosaic. Annals of Applied Biology 106, 285-294.
Fargette, D., M. Jeger, C. Fauquet, and L.D.C. Fishpool. 1994. Analysis of temporal 
disease progress of African cassava mosaic virus. Phytopathology 84:91-98.
Fargette, D. and K. Vie, 1988. Yield losses induced by African cassava mosaic virus in 
relation to mode and date of infection. Tropical Pest Management 34:89-91.
Fargette, D. and K. Vie, 1994. Modelling the temporal primary spread of African 
cassava mosaic virus into plantings. Phytopathology 84: 378 -  382.
Fargette, D. and K. Vie, 1995. Simulation of the effects of host resistance, reversion and 
cutting selection on incidence of cassava mosaic virus and yield losses in 
cassava. Phytopathology 85: 370-375.
Fauquet, C and R.N., Beachy, 1991. Cassava virus and genetic engineering. 
International Cassava Trans Project (CTA) 30 pg.
183
Fauquet, C. and D. Fargette, 1990. African cassava mosaic virus: aetiology, 
epidermiology and control. Plant Diseases 74(6): 404-411.
Fauquet, C., D. Fargette and J.C. Thouvenel, 1987a. Selection of healthy cassava 
plants obtained by reversion in cassava fields. In; The International Seminar on 
African Mosaic Disease and its Control, Yamoussoukro, Cote d’Ivoire 4-8 May, 
1987, . pg. 146-149. CTA, FAO, ORSTOM, IITA, IAPC publication. Wageningen 
The Netherlands.
Fauquet, C., D. Fargette and J.C. Thouvenel, 1987b. The resistance of cassava to 
African cassava mosaic. In: The International Seminar on African Mosaic 
Disease and its Control, Yamoussoukro, Cote d’Ivoire 4-8 May 1987. pg 183- 
188. CTA, FAO, ORSTOM, IITA, IAPC publication. Wageningen, The 
Netherlands.
Fauquet, C., D. Fargette and J.C. Thouvenel, 1988. Some aspects of the epidermiology 
of African cassava mosaic virus in Ivory Coast. Tropical Pest Management 34: 
92-96.
FensterC.B. and K. Ritland, 1994. Quantitative genetics of mating system divergence in 
the yellow monkeyflower species complex. Heredity 73:422-435.
Feyt, H., 1976. Etude critque de I’analyse des croisements dialles au moyen de la 
simulation. Annales de lAmelioration des Plantes 26:173-193.
Fondong, V.N., F.S. Pita, M.E.C. Rey, A. de Kochko, R.N. Beachy and C.M. Fauquet, 
2000. Evidence of synergism between African cassava mosaic virus and a new 
double-recombinant geminivirus irifecting cassava in Cameroon. Journal of 
General Virology 81:287-297.
Fontes, E.P.B, V.A Luckow, L. Hanley-Bowdoin 1992) A geminivirus replication protein 
is a sequence specific DNA binding protein. Plant Cell 4:597-608.
Fox, S.L., P.D. Brown and J. Chong, 1997. Inheritance of crown rust resistance in four 
accessions of Avena sterilis L. Crop Science 37:342-345
Fraser, R.S.S., 1990. Disease resistance mechanisms. In: C.L. Mandachar and C.L. 
Mandachar (eds) Plant Viruses Volume II. Pathology, pg 321-345, CRC Press, 
Boca Raton, FL, USA.
Fregene, M.A., 1996. Phylogeny of cassava (Manihot esculenta Crantz) and its wild 
relatives based on restriction fragment length polymorphism (RFLP) analysis. 
PhD Thesis University of Ibadan, Ibadan Nigeria.
Fregene, M., F. Angel, R. Gomez, F. Rodriguez, P. Chavarriaga, M. Bonierbale, W. 
Roca and J. Tohme, 1997. A molecular genetic map of cassava (Manihot 
esculenta Crantz). Theoretical and Applied Genetics 95:431-441.
Fregene, M.A., F. Angel, R. Gomez, F. Rodriguez, M. Mayer, J. Tohme, M. Bonierbale, 
C. Iglesias, and W.M. Roca, 1994. A linkage map of cassava (Manihot 
esculenta Crantz) based on RFLP and RADP markers. In: International 
Cassava Biotechnology Network: Proceedings of the 2nd. International 
Scientific. Meeting, Bogor Indonesia 22-26 August 1994.
184
Fregene, M., A. Bernal, M. Duque, A. Dixon and J. Tohme, 2000. AFLP analysis of 
African cassava (Manihot esculenta Crantz) germplasm resistant to the cassava 
mosaic disease (CMD). Theoretical and Applied Genetics 100: 678-685
Fregene, M., E. Okgbenin, C. Mba, F. Angel, M.C. Suarez, J. Guitierez, P. Chavarriaga, 
W. Roca, M. Bonierbale and J. Tohme, 2001. Genome mapping in cassava 
improvement:: Challenges, achievements and opportunities. Euphytica 120: 
159-165.
Fritz, A.K., S. Caldwell and W.D. Worall, 1999. Molecular mapping of Russian wheat 
aphid resistance from triticale accession PI38156. Crop Science 39:1707-1710.
Gale, M.D and K.M. Devos, 1998. Plant comparative genetics after 10 years. Science 
282:656-659.
Garcia, G.M., H.T., Stalker, E. Shroeder and G. Kochert, 1996. Identification of RAPD, 
SCAR and RFLP markers tightly linked to nematode resistance genes 
introgressed from Arachis cardenasii into Arachis hypogaea. Genome 39:836- 
845.
Gardener, C.O. and S.A. Eberhart, 1966. Biometrics 22(3):439-452.
Geddes, A. M., 1990. The relative importance of crop pests in sub-Saharan Africa. 
Bulletin 36; Chatham Natural Resource Institute (NRI).
Gentzbittel L., S. Mouzeyar, S. Badaoui, E. Mestries, F. Vear, D.T. de Labrouhe, P. 
Nicolas, 1998. Cloning of molecular markers for disease resistance in sunflower, 
Helianthus annuus L. Theoretical and Applied Genetics 96:519-525.
Gibson,. R.W., 1994. Long-term absence of symptoms in heat-treated African cassava 
mosaic geminivirus infected resistant cassava plants. Tropical Science 34:154- 
158.
Gilbert, N.E., 1973. Biometrical Interpretation. Clarendon Press, Oxford, pg 125.
Giovannoni, J.J., R.A. Wing, M.W. Ganal,. S.D. Tanksley, 1991. Isolation of molecular 
markers from specific chromosomal intervals using DNA pools from existing 
mapping populations. Nucleic Acids Research 19(23):6553-6558.
Gomez, R. F. Angel, M.W. Bonierbale, F. Rodriguez, J. Tohme and W.M. Roca, 1996. 
Inheritance of random amplified polymorphic DNA markers in cassava (Manihot 
esculenta Crantz) Genome 39:1039-1043.
Grattapagalia, D., C. Costa de Silva, and N.M.A. Nassar, 1996. Strict apomix, in 
cassava (Manihot esculenta Crantz). Canadian Journal of Plant Science 76:379- 
382
Gravois, K.A., 1994. Diallel analysis of head rice percentage, total milled rice 
percentage, and rough rice yield. Crop Science 34:42-45.
Griffing, B., 1956a. A generalized treatment of the use of diallel crosses in quantitative 
inheritance. Heredity 10: 31-50.
Griffing, B., 1956b. Concept of general and specific combining ability in relation to diallel 
crossing systems. Australian Journal of Biological Science 9: 463-493.
185
Gulick, P., C. Hershey and J. Esquinas-Alcazar, 1983. Genetic Resources of cassava 
and wild relatives. IBPGR report 82/III. International Board for plant genetic 
resource, Rome Italy.
Hahn, S.K. 1978. Breeding cassava for resistance to cassava mosaic disease and 
bacterial blight in Africa. In: H. Maraite and Y.A. Meyer (Eds), Diseases of 
Tropical Food Crops, Proceedings. International Symposium at the University 
Catholique Louvain pg 211-219. Louvain La Neuve, Belgium.
Hahn, S.K. and A.K. Howland, 1972. Breeding for resistance to cassava mosaic 
disease. In: Terry, E.R (Ed) Proceedings, Cassava mosaic workshop, Ibadan 
Nigeria pg 37-39. IITA, Ibadan Nigeria.
Hahn, S.K., A.K. Howland and E. R. Terry, 1980a. Correlated resistance of cassava to 
mosaic and bacterial blight disease. Euphytica 29:305-311
Hahn S.K., C. John, G. Isoba and T. Ikoun, 1989. Resistance breeding in root and tuber 
crops at the International Institute for Tropical Agriculture (IITA), Ibadan Nigeria. 
Crop Protection 8:147-168.
Hahn, S.K., E.R. Terry and K. Leuschner, 1980b. Breeding cassava for resistance to 
cassava mosaic virus disease. Euphytica 29:673-683
Hahn, S. K., E.R. Terry, K. Leuschner, I.O. Akobundu, C. Okali and R. Lai, 1979. 
Cassava improvement in Africa. Field Crops Research. 2:193-226.
Haley, S.D., L.K. Afanador, P.N. Miklas, J.R. Stavely and J.D. Kelly. 1994. 
Heterogeneous inbred populations are useful as source of near-isogenic lines 
for RAPD marker localisation. Theoretical and Applied Genetics 88:337-342.
Haley, S.D., P.N. Miklas, J.R. Stavely, J. Byrum and J.D. Kelly, 1993. Identification of 
RAPD markers linked to a major rust resistance gene block in common bean. 
Theoretical and Applied Genetics 86:505-512.
Hallauer A.R. and J.B. Miranda, 1988. Quantitative genetics in maize breeding. 2nd Ed 
Iowa State University 468pg. Ames, USA.
Hamada, H., M.G. Pertino, and T. Kakunaga. 1982. A novel repeated element with Z- 
DNA- forming potential is widely found in evolutionary diverse eukaryotic 
genomes. Proceedings, National Academy of Science (USA) 79:6465-6469.
Harrison, B.D., Y.L. Liu, X. Zhou, D.J. Robinson, L. Calvert, C. Munoz and G.W. Otim- 
Nape, 1997b. Properties, differentiation and geographical distribution of 
geminiviruses that cause cassava mosaic disease. African Journal of Root and 
Tuber Crops 2:19-22.
Harrison, B.D. and D.J. Robinson, 1988. Molecular variation in vector-borne plant 
viruses: epidemiological significance. Philosophical Transactions of the Royal 
Society of London 321:447-462.
Harrison, B.D., X. Zhou, G.W. Otim-Nape, Y. Liu and D.J. Robinson, 1997a. Role of a 
novel type of double infection in the geminivirus-induced epidemic of severe 
cassava mosaic in Uganda. Annals of Applied Biology 131:437-448.
186
Hauge, B.M, S.M. Haney, S. Cartinhour, J.M. Cherry, H.M. Goodman, M. Koornneof, P. 
Stam, C. Chang, S. Kempin, L. Medrano and E.M. Meyerowitz, 1993. An 
intergrated genetic /RFLP map of Arabidopsis thaliana genome. The Plant 
Journal 3(5):745-754.
Hayman, B.I., 1954a. The analysis of variance of diallel tables. Biometrics 10: 235-244.
Hayman, B.I., 1954b. The theory and analysis of diallel crosses. Genetics 39: 789-809
Haysom, H.R., T.L.C. Chan and M.A. Hughes, 1994. Phylogenetic relationships of 
Manihot species revealed by restriction fragment length polymorphism. 
Euphytica 76:277-234.
Hershey C., 1987. Cassava germplam resources. Proceedings workshop on cassava 
breeding: A multidisciplinary review. The Philippines, 4-7 March 1985. CIAT, 
Cali Colombia.
Hoefer Scientific, 1992 Operating instructions TKO 100 minifluorometer. Hoefer 
Scientific Instruments, San Francisco CA., USA. 50pg.
Holland, J.B., 2001. Epistasis and plant breeding. In: Janick, J. (Ed). Plant Breeding 
Reveiws. 21:27-91.
Holloway, J. and S. Knapp, 1994. GMendel 3.0.Users Guide, knapps@css.orst.edu
Hong, Y.G., D.J. Robinson and B.D. Harrison, 1993. Nucleotide sequence evidence for 
the occurrence of three distinct whitefly-transmitted geminiviruses in cassava. 
Journal of General Virology 74:2437-2443.
Horvat, F. and M. Verhoyen, 1981. Cytological modifications and presence of virus-like 
particles in cells of Nicotinana benthamiana (Domin) and Manihot utilissima 
(Pohl.) infected with the geminivirus isolated from cassava infected with cassava 
African mosaic disease. Parasitica 37:121-130.
Hussain, A., W. Bushuk., H. Ramirez and W. Roca, 1987. Identification of cassava 
(Manihot esculenta Crantz) cultivars by electrophoretic patterns of esterase 
enzymes. Seed Science Technolgy 15:19-21.
Iglesias C.A and C.H. Hershey, 1991. True cassava seed: research for a production 
alternative. In: Ofori, F and S.K. Hahn (Eds).Tropical root crops in a developing 
economy. Proceedings 9th symposium of the international society of tropical 
crops 20-26 Oct 1991, Accra, Ghana, pg. 164-171. Wageningen, The 
Netherlands.
IITA, 1973. In house report 1972. Root and tuber crop improvement programme. 
International Institute for Tropical Agriculture, Ibadan Nigeria pg 17-23.
IITA, 1988,1980,1997, 1998, 2000. Annual Report of the International Institute of 
Tropical Agriculture. IITA Ibadan Nigeria.
IITA, 1990. Cassava in tropical Africa. A reference manual. Chayce Publication 
Services, United Kinqdom Edition. Baling & Mansell International, Wisbech UK. 
196 pg.
187
Janessen, P., R. Coopman, G. Huys, J. Swings, M. Bleeker, P. Voss, M, Zabeau and K. 
Kersters, 1996. Evaluation of the DNA fingerprinting method AFLP as a new tool 
in bacteria taxonomy. Microbiology 142:1881-1893.
Jeffrey, A.J., V. Wilson, and S.L. Thein. 1985. Hypervariable minisatellite regions in 
human DNA. Nature 314:67-73.
Jeger, M.J and S.L.H. Viljanen-Rollinson, 2001. The use of area under the disease 
progress curve (AUDPC) to assess quantitative disease resistance in crop 
cultivars. Theoretical and Applied Genetics 102:32-40.
Jennings, D.L., 1960. Observations on virus diseases of cassava in resistant and 
susceptible varieties: I Mosaic disease. Empire Journal of Experimental 
Agriculture 28: 23-34.
Jennings, D.L., 1977. Inheritance of linked resistance to African cassava mosaic and 
bacteria blight. In: Proceedings Cassava Protection workshop. CIAT. Colombia 
pg 45-50.
Jennings, D.L., 1976. Breeding for resistance to African cassava mosaic. In: African 
cassava mosaic Report of an interdisciplinary workshop held at Muguga, Kenya, 
pg. 60-87. IITA, Ibadan, Nigeria
Jennings, D.L., 1994. Breeding for resistance to African cassava mosaic virus in East 
Africa. Tropical Science 34:110-122.
Jennings, D.L. and C.H. Hershey, 1985. Cassava breeding: a decade of progress from 
international programmes. In; Russell G.E. (Ed.) Progress in Plant Breeding I. 
Butterworths, Cambridge pg 89-116.
Jinks, J.L. and B.l. Hayman, 1953. The analysis of diallel crosses. Maize Genetics 
Cooperative Newsletter 27: 48-54.
Jos, J.S., and N.G. Nair, 1979. Pachytene pairing in relation to pollen fertility in five 
cultivars of cassava. Cytologia (Tokyo) 44:813-820.
Jos, J.S., K.V. Bai and N.G. Nair, 1984. Unreported symptoms associated with mosaic 
disease in cassava. Indian Phytopathology 37:696-699.
Kaiser, W.J. and L.R. Teemba, 1979. Use of tissue culture and thermotherapy to free 
East African cassava cultivars of African cassava mosaic and cassava brown 
streak disease. Plant Disease Reporter 63:780-784.
Kaiser, W.J. and R. Louie, 1982. Heat therapy of cassava infected with African cassava 
mosaic disease. Plant Disease 66: 475-477.
Kanazin, V., L.F. Marek, and R.C. Shoemaker, 1996. Resistance gene analogs are 
conserved and clustered in soybean. Proceedings of National Academy of 
Science (USA) 93:11746-11750.
Kang, M.S. 1994. Applied quantitative genetics. 2477 Creekside Drive Baton Rouge. 96 
pg. Baton Rogue, FL., USA.
Kang, M.S., 1998. Using Genotype by Environment interaction for crop cultivar 
development. Advances in Agronomy, 62:199-258.
188
Kang, M.S, A. Kushairi, Y. Zhang and R. Magari, 1999. Combining ability for rind 
puncture resistance in maize. Crop Science 39:368-371.
Kawano, K, 1978. Genetic improvement of cassava (Manihot esculenta Crantz) for 
productivity. Tropical Agriculture Research Series. II. Ministry of Agriculture and 
Forestry, Japan. 21 pg.
Kawano, K, 1980. Cassava. In Hybridisation of crop plants. American Society of 
Agronomy and Crop Science Society of America, pg 255-233. Madison, 
Wisconsin. USA.
Kelly, J.D., 1995. Use of random amplified polymorphic DNA markers in breeding for 
major gene resistance to plant pathogens. HortScience 30: 461-465.
Kempthorne, O. 1956. The theory of the diallel cross. Genetics, 41: 451-459.
Kempthorne, O. and R.N. Curnow, 1961. The partial diallel cross. Biometrics, 17: 229- 
250.
Kiehne, K, and D.B. Neale, 1998. DNA pooling strategy for saturation mapping in 
outcrosses. Molecular Breeding 4:179-185.
Knapp, S.J., W.W. Stoup and W.M. Ross, 1985. Exact confidence intervals for 
heritability on a progeny means basis. Crop Science 25:192-194.
Kosambi, D. D. 1944. The estimation of map distances from recombination values. 
Annals of Eugenics 12:172-175.
Kwok, S. D.E. Kellogg, N. McKinney, D. Spasic, L. Goda, C. Levenson and J.J. 
Sninsky, 1990. Effects of primer-template mismatches on the polymerase chain 
reaction: human immunodeficiency virus type 1 model studies. Nucleic Acids 
Research 18 (4): 999-1005.
Laminski, S., E.R. Robinson and V.M. Gray, 1997: Application of molecular markers to 
describe South African elite cassava cultivars. African Journal of Root and 
Tuber Crops 2:132-134.
Lancaster, P.A. and J.E. Brooks, 1983. Cassava leaves as human food. Economic 
Botany 37(3):33-348.
Lande, R., 1981. Minimum number of genes contributing to quantitative variation 
between and within populations. Genetics 99:541-553.
Lander, E.S. and D. Botstein, 1986. Mapping complex genetic traits in humans: New 
method using a complex RFLP linkage map Cold Spring Harbor Symposia on 
Quantitative Biology 51: 49-62.
Lander, E.S., P. Green, J. Abrahamson, A. Barlow, M.J. Daly, S.E. Lincoln, and L. 
Newberg, 1987. MAPMAKER: An interactive computer package for constructing 
primary genetic linkage maps of experimental and natural populations. 
Genomics 1:174-181.
Lawrence, P.L and K.J. Frey, 1976. Inheritance of grain yield in oats species crosses 
Avena sativa L. x A. sterilis L. Egyptian Journal of Genetics and Cytology 5:400- 
409.
189
Lefevre, F. and A. Charrier, 1993. Isozyme diversity within African Manihot germplasm. 
Euphytica 66:73-80.
Legg, J .P ., 1994. Bemisia tabaci: T he  w h ite fly  v e c to r  o f c a ssa va  m osa ic  g em in iv iru se s  
in A fr ic a : A n  eco log ica l pe rspec tive . African Crop Science Journal 2: 437-448.
Leonards-Schippers, C., W. Gieffers, R. Schaufe-Pregl, E. Ritter, S.J. Knapp, F. 
Saamini and C. Gebhardt, 1994. Quantitative resistance to Phytphthora 
infestans in potato: a case study of QTL mapping in an allogamous plant 
species. Genetics 137:67-77.
Leuschner, K., 1978. Whiteflies: biology and transmission of African cassava mosaic 
disease. In. T. Brekelbaum, A. Belloti and T.C. Lozano (Eds.). Proceedings of 
the Cassava Protection Workshop, CIAT, Cali, Colombia 7-12 November 1977, 
pg 51-58. CIAT, Colombia;
Littell, R.C., G.A. Miliken, W.W. Stoup and R.D. Wolfinger, 1996. SAS system for mixed 
models. SAS Institute Inc., Cary, North Carolina, USA. 633 pg.
Liu, B. H., 1997. Statistical Genomics: Linkage mapping and QTL analysis. CRC Press 
LLC, 611 pg. Boca Raton, Florida, USA.
Liu, B. H., 1998. Computational tools for study of complex traits. In A.H. Paterson (Ed) 
Molecular dissection of complex traits. CRC Press LLC pg 43-79. Boca Raton, 
Florida, USA
Lohmann, S. L. Lehmann and K. Tahiti, 2000. Fast and flexible single nucleotide 
polymorphism (SNP) dtection with the LightCycier System. Biochemica 4:23:28.
Lokko, Y., A.G.O. Dixon and S.K. Offei, 1998. Combining Ability of resistance to the 
cassava mosaic virus disease. In M.O. Akorada and J.M. Ngeve (Eds). Root 
Crops in the 21st Century. Proceedings of the 7th Triennial conference of the 
International Society for Tuber and Root Crops, Africa Branch,. 11-17 Oct. 
1998 Cotonou, Benin, pg 438-442. IITA Ibadan, Nigeria.
Lozano, J.C. 1992. Overview of integrated control of cassava diseases. Fitopatologia 
Brasileira 17:18-22.
Lynch, M. and B. Walsh, 1998. Genetic Analysis of Quantitative traits. Sinauer 
Associates Inc., Sunderland, MA., USA 985 pg.
Magoon, M.L. 1967. Recent trends in cassava breeding in India. In: University of west 
Indies. In; E.A. Tai, W.B. Charles, P.H. Haynes, E.F. Iton, K.A. Leslie (Eds) 
Proceedings first International Symposium of Tropical Root Crops, held at the 
University of West Indies, vol 2 pg 100-117 St. Augustine, Trinidad.
Magoon, M.L., and R.K., Krishnan, 1977. Extending the frontiers of genetic 
improvement in cassava. In: Leaky L.A. (ed.). Proceedings third symposium of 
the international society for tropical root crops. IITA Ibadan, Nigeria, 2-9 Th. Dec 
1973, pg. 14-18.
Magoon, M.L., R. Krishnan and K.V. Bai, 1969. Morphology of pachytene chromosome 
and meiosis in Manihot esculenta Crantz. Cytologia 34:612-626.
190
Maluszynski, M., B.S. Ahloowalia, B. Sigurbjornsson, 1995. Application of in vivo and in 
vitro mutation techniques for crop improvement. Euphytica 85.303-315.
Mandal, R.C., 1993. Tropical root and tuber crops. Cassava (tropical), sweetpotato, 
aroids! yam yam beans and coleus. Agrobotanicai Publishers pg 130-133. 
Bikane, India
Manly, K.F. and E.H. Cudmore Jr., 1996. New versions of MapManager genetic 
mapping software. Plant genome IV (Abs) pg105.
Marmey, P. J.R., Beeching., S. Hammon and A. Charrier, 1994. Evaluation of cassava 
(Manihot esculenta Crantz) germplasm using RAPD markers. Euphytica 
74:203-209.
Mather, K. and J.L. Jinks, 1982. Biometrical Genetics: The Study of Continuous 
Variation. Longman, London. 396 pg.
Matzinger, D.F. and O. Kempthorne, 1956. The modified diallel table with partial 
inbreeding and interactions with environment. Genetics 41:822-833.
Mba, R.E.C, P. Stephenson, K. Edwards, S. Melzer, J. Nkumbira, U. Gullberg, K. Apel, 
M. Gale, J. Tohme and M. Fregene, 2001, Simple sequence repeat (SSR) 
markers survey of the cassava (Manihot esculenta Crantz) genome: towards an 
SSR-based molecular genetic map of cassava Theoretical and Applied 
Genetics 102:21-31.
McIntosh, M.S., 1983. Analysis of combined experiments. Agronomy Journal 75:153- 
155.
McMullen, MD, M.W. Jones, K.D. Simcox and R. Louie. 1994 Three genes control 
resistance to wheat streak mosaic virus in the maize inbred Pa405. Molecular 
plant. Microbe Interaction 7:708-712.
McMullen, M.D. and R. Louie. 1989. Linkage of molecular markers to a gene controlling 
the symptom response in maize to maize dwarf mosaic virus. Molecular plant. 
Microbe Interaction 2(6):309-314.
Melchinger, A.E. 1990. Use of molecular markers in breeding for oligogenic disease 
resistance. Plant Breeding 104:1-19.
Melchinger, A.E., L. Kuntze, R. K. Gumber, T.L Ubberstedt and E. Fuchs 1998. Genetic 
basis of resistance to sugarcane mosaic virus in European maize germplasm 
Theoretical and Applied Genetics 96:1151-1161.
Michelmore, R.W, 1995 Molecular approaches to manipulation of disease resistance 
genes. Annual Review of Phytopathopathology 15:393-427. ;
Michelmore, R. W., I. Paran and R.V. Kesseli, 1991. Identification of markers linked to 
disease-resistant genes by bulked segregant analysis: A rapid method to detect 
markers in specific genomic regions by using segregating populations. 
Proceedings of National Academy of Science (USA) 88:9828-9832.
Mignouna, H.D. and A.G.O Dixon, 1997. Genetic relationships among cassava clones 
with varying levels of resistance to the African mosaic disease using RAPD 
markers. African Journal of Root and Tuber Crops 2 (1&2):28-32.
191
Miklas, P.N., L. Afanadorand J.D. Kelly, 1996. Recombination-facilitated RAPD marker- 
assisted selection for disease resistance in common bean. Crop Science 36:86- 
90.
Miklas, P.N., J.R. Stavely, and J.D. Kelly, 1993. Identification and potential use of a 
molecular markers for rust resistance in common bean. Theoretical and Applied 
Genetics 85:745-749.
Moll, R.H., W.S. Salhuana and H.F. Robinson, 1962. Heterosis and genetic diversity in 
variety crosses of maize. Crop Science 2:197-198.
Morgante, M. and A. M. Olivieri, 1993. PCR-amplified microsatellites as markers in plant 
genetics. Plant Journal 1:175-182.
Msabaha, M.A.M., 1983. Genetic relationship between resistance to mosaic and 
bacterial blight disease, in cassava. (M. esculenta Crantz). PhD Thesis 
University of Ibadan. 215 pg.
Msikita, W., B. James, E. Nnodu, J. Legg, K. Wydra and F. Ogbe, 2000 Disease Control 
in Cassava Farms I PM Field Guide for Extension Agents. International Institute 
of Tropical Agriculture, Ibadan Nigeria. 26 pg.
Muehlbauer, G.J, J.E., Specht, M.A. Thomas-Compoton, P.E, Staswick, and R.L 
Bernard, 1988. Near-isogenic lines-A potential resource in the integration of 
conventional and molecular marker linkage maps. Crop Science 28:729-735.
Multize, D.K. and R.J. Barker, 1985. Evaluation of biometrical methods for estimating 
the number of genes. Theoretical and Applied Genetics 69:559-566.
Nassar, N.M.A., 1994. Development and selection of apomixis in cassava. Canadian 
Journal of Plant Science 74(4)857-858.
Ng, S.Y.C., G. Thottappilly and H.W. Rossel, 1992. Tissue culture in disease elimination 
and micropropagation. In: G. Thottappilly, L.M. Monti, D.R. Mohan Raj, A.W. 
Moore (eds.) Biotechnology: Enhancing Research on Tropical Crops in Africa, 
pg 171-182, CTA/IITA co-publication. IITA, Ibadan, Nigeria.
Nguyen, H.T. and D.A Sleper, 1983. Genetic variability of seed yield and reproductive 
characters in tall fescue. Crop Science 23:621-626.
Nian C.A., S. Abraham, L.A. Bia and A. Kariachen, 1976. Lipid metabolism in cassava 
mosaic infected Manihot esculenta Crantz. 4th Symposium. Tropical Root 
Crops. CIAT, Cali Colombia 1976. Cali, Colombia.
Nichols, R.F.W., 1947. Breeding cassava for virus resistance. East African Agricultural 
Journal 15:154-160.
Njock, T.E., 1994. Epidemiology of disease recovery phenomenon of ACMV in resistant 
and susceptible cassava clones. PhD Thesis, University of Ibadan, Ibadan, 
Nigeria. 194pg.
Njock, T.E., G.l. Atiri, G. Thottappilly and J.M. Thresh, 1996. Disease development and 
recovery in resistant and susceptible cassava genotypes infected by African 
cassava mosaic geminivirus. African Plant Protection 2:97-102.
192
Njock, T.E., G.l. Atiri, G. Thottappilly and S.K. Hahn, 1994. Distribution and disease 
recovery of African cassava mosaic virus (ACMV) in stems of resistant and 
susceptible cassava. In: Proceedings of the fifth Triennial Symposium of the 
International Society of Tropical Root Crop, Africa Branch (iSTRC-AB), pg 149- 
152. IITA, Ibadan, Nigeria.
Ocampo, C., C.H. Hershey C.A. Iglesias and M. Iwansa, 1993. Esterase Isozyme 
fingerprinting of CIAT cassava germplasm collection In: Roca, W.M., and A.M. 
Thro (Eds), Proceedings of the first international meeting of the cassava 
biotechnology network. Cartagena de Indias, Colombia, 81-89 August, 1992 pg 
51-55. Working document No. 123. CIAT, Cali, Colombia.
Offei, S. K., M. Owuna-Kwakye and G. Thottappilly, 1999. First report of East African 
cassava mosaic begomovirus in Ghana. Plant Disease 83:877.
Ogbe, F.O., 2001. Survey of cassava begmoviruses in Nigeria and the response of 
resistant cassava genotypes to African cassava mosaic begmovirus infection. 
PhD Thesis University of Ibadan, Nigeria. 197 pg.
Ogbe, F.O., G.l. Atiri, D. Robinson, A.G.O. Dixon, S. Winter, F.M. Quin and G. 
Thottappilly, 1999. First report of East African cassava mosaic begomovirus in 
Nigeria. Plant Disease 83(4):398.
Ogbe, F.O., E.C. Nnodu and S.O, Odurukwe, 1996. Control of African cassava mosaic 
disease incidence and severity. Tropical Science 36(2):174-181.
Ori, N., Y. Eshed, I. Paran, G. Presting, D. Aviv, S. Tanksley, D. Zamir, and R. Fluhr, 
1997. The I2C family from wilt disease resistance locus 12 belongs to the 
nucleotide binding, leucine-rich repeat superfamily of plant resistance genes. 
The Plant Cell 9:521-532.
Oslen, K.M. and B.A. Schaal 1999. Evidence on the origin of cassava: Phylogeography 
of Manihot esculenta. Proceedings of the National Academy of Science USA 
96:5586-5591.
Otim-Nape, G.W., A. Bua and Y. Baguma, 1994b. Accelerating the transfer of improved 
production technologies: Controlling African cassava mosaic virus disease 
epidemic in Uganda. African Crop Science Journal 2: 479-495.
Otim-Nape, G.W., J. Shaw and M. Thresh, 1994a. Effects of African cassava mosaic 
geminivirus on the growth and yield of cassava in Uganda. Tropical Science 
34:43-54.
Otim-Nape, G.W., J.M. Thresh, A. Bua, Y. Baguma and M.W. Shaw, 1998. Temporal 
spread of cassava mosaic virus disease in a range of cassava cultivars in 
different agro-ecological regions of Uganda. Annals of Applied Biology 133:415- 
430.
Paterson, A., J. DeVerna, B. Lanini, and S. Tanksley. 1990. Fine mapping of 
quantitative trait loci using selected overlapping recombinant chromosomes, in 
an interspecies cross of tomato. Genetics 124:735-742.
Paterson, A.H, E.S. Lander, J.D., Hewitt, S. Peterson, S.E. Lincoln and S.D. Tanksley. 
1988. Resolution of quantitative trait into mendelian factors by using a complete 
RFLP linkage map. Nature 335:721-726.
193
Pedersen, J.F., F.A. Haskins, H.J., Gorz and R. Brittons, 1983. Variability for traits used 
to estimate silage quality in forage sorghum hybrids. Crop Science 23:376-379.
Pillay, M., and S.T., Kenny, 1996. Random amplified polymorphic DNA (RAPD) markers 
in’ hops, Humulus lupulus: level of genetic variability and segregation in F-, 
progeny. Theoretical and Applied Genetics 92:334-339.
Pooni, H.S., J.L., Jinks, and R.K. Singh, 1984. Methods of analysis and the estimation 
of the genetic parameters from a diallel set of crosses. Heredity 52 (2):243-253.
Porto, M.C., R. Asiedu, A.G.O. Dixon, and S.K. Hahn, 1994. An agroecological oriented 
introduction of cassava germplasm from Latin America into Africa. In: Ofori, F 
and S.K. Hahn (Eds).Tropical root crops in a developing economy. Proceedings 
9th symposium of the international society of tropical crops 20-26 Oct 1991, 
Accra, Ghana. Wageningen, The Netherlands.
Powell, W., G.C. Machray and J. Provan, 1996b Polymorphism revealed by simple 
sequence repeats. Trends in Plant Science 1:215-222.
Powell, W., M. Morgante, C. Andre, M. Hansafey, J. Vogel, S. Tingey and A. Rafalski, 
1996a. The comparison of RFLP, RAPD, AFLP and SSR (microsatellite) 
markers for germplasm analysis. Molecular Breeding 2:225-238.
Promega, 2000. Technical manual silver staining™ DNA sequencing system. Promega 
Corporation, Madison, Wl. USA. 24 pg.
Qiagen, 1999. DNeasy ™ Plant Mini Handbook for DNA isolation from plant tissue. 
QIAGEN Ltd. Crawley, West Sussex, UK. 22 pg.
Rajendran, P.G., 1989. Combining ability in cassava. Journal of Root Crops. 15(1):15- 
18..
Raji, A.A., 1995. Evaluation of pest and disease reactions and trait association in some 
landraces of cassava. MSc Thesis, University of Ibadan, Ibadan, Nigeria, 86 pg.
Ramirez, H., A. Hussain, W. Roca and W. Bushuk, 1987. Isoenzyme electrophorogram 
of sixteen enzymes in five tissues of cassava (Manihot esculenta Crantz) 
varieties. Euphytica 36: 39-48.
Reamon-Buttner, S.M., J. Schondelmaier and C. Jung, 1998. AFLP markers tightly 
linked to the sex locus in Asparagus officinalis L. Molecular Breeding 4:91-98.
Rhode, W. 1996. Inverse sequence-tagged repeat (ISTR) analysis a novel and 
universal PCR-based technique for genome analysis in the plant and animal
kingdom. Journal of Genetics &. Breeding 50:249-261.
Rick, C.M., and P.G. Smith, 1953. Novel variation in tomato species hybrids. The 
American Naturalist 88:359-373.
Roa, A., P. Chavarriaga-Aguirre, M.C. Durque, M.M. Maya, M.W. Bonierbale, C. 
Iglesias and J. Tohme, 2000. Cross-species amplification of cassava (Manihot 
esculenta) (Euphorbiaceae) microsatellites: Allelic polymorphism and degree of 
relationship. American Journal of Botany 87(11):1647-1655.
194
Roa, A., MM  Maya M Durque, M., J. Tohme, A. Allem, and M. Bonierbale, 1997. 
AFLP analysis of relationships among cassava and other Manihot species. 
Theoretical Applied Genetics 95:741-750
Robinson, D.J., B.D. Harrison, J.C. Sequeira, and G.H. Duncan, 1984. Detection of 
strains of African cassava mosaic virus by nuclei acid hybridisation and some 
effects of temperature on their multiplication. Annals of Applied Biology 105:483- 
493.
Rooney, W.L., D.Z. Skinner and J.O. Fritz, 1997. Combining ability for protein 
degradability in Alfalfa. Crop Science 37:128-131.
Ross, W.M, H.J. Gorz, F.A. Haskins, G.H. Hookstra, J.K. Rutto and R. Ritter, 1983. 
Combining ability effects for forage residue traits in grain sorghum hybrids. Crop 
Science 23:97-101.
Rossel, H.W., R. Asiedu, and A.G.O. Dixon 1992. Resistance of cassava to African 
cassava mosaic virus: What really pertains. Tropical Root Tuber Crops Bulletin 
6:2.
Rusell, J.R., J.D. Fuller, M. Macaulay, B.G. Hatz, A. Jahoor, W. Powell and R. Waugh,
1997. Direct comparison levels of genetic variation among barley accessions 
detected by RFLP’s, AFLP’s, SSR’s and RAPD. Theoretical and Applied 
Genetics 95:714-722.
Sambrook, J., E.F. Fritsch, and T. Maniatis, 1989. Molecular cloning: a laboratory 
manual. 2nd ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, USA.
Sanchez G., S. Restrepo, M. Duque, M. Fregene, M.W. Bonierbale and V. Verdier, 
1999. AFLP assessment of genetic variability in cassava accessions (Manihot 
esculenta Crantz) resistant and susceptible to the cassava bacterial blight (CBB) 
Genome 42:163-172.
SAS, 1997. SAS/STAT Software. Changes and enhancements through Release 
version 6.12. SAS institute, Cary, North Carolina, USA. pg 1028.
SAS Institute, 1999. SAS companion for Microsoft windows environment, version 6, 1st 
ed. SAS Institute, Cary, North Carolina, USA.
Sax, K. 1923 The association of size differences with seed-coat patterns and 
pigmentation in Phaseolus vulgaris. Genetics 8:552-560.
Schaal, B., L.J.C.B. Carvalho, T. Prinzie, K. Olsen, P. Olson, G. Cabral and M. 
Hernandez, 1997. Phylogenetic relationships among Manihot species. Africa. 
African Journal of Root and Tuber Crops 2:147-149.
Schopke, C., R. Carcano, R.N. Beachy and C. Fauquet, 1997. Plant regeneration from 
transgenic and non-transgenic suspension cultures of cassava (M. esculenta 
Crantz). African Journal of Root and Tuber Crops 2:194-195.
Second, G., A.C. Allem, L. Emperarie, C. Ingram, C. Columbo, R.A. Mendes and L.J. 
C.B. Carvalho, 1997. AFLP based Manihot and cassava numerical taxonomy 
and genetic structure analysis in progress: Implications for dynamic 
conservation and genetic mapping. African Journal of Root and Tuber Crops 
2:140-146.
195
Seif, A.A., 1982. Effect of cassava mosaic virus on yield of cassava. Plant Disease 
66:661-662.
Seo, Y.W., J.W. Johnson, and R.L. Jarret, 1997. A molecular marker associated with 
H21 Hessian fly resistance gene in wheat. Molecular Breeding 3:177-181.
Shen, K.A., B.C. Meyers, M.N. Islam-Faridi, D.B. Chin, D.M. Stelly, R.W. Michelmore,
1998. Resistance genes candidates identified by PCR with degenerate 
oligonucleotide primers map to clusters of resistance genes in lettuce. Molecular 
Plant Microbe Interaction 11:815-823.
Singh, R.K. and B.D. Chaudhary, 1985. Biometrical methods in quantitative genetics 
analysis. Kalyan Publications 3rd Ed. Pg 205-220.
Silvestre, P. and M. Arraudeaus, 1983. Le manioc. ACCT, coll. Techniques, Agricoles et 
Productions Tropicales. Maisonneuve & Larose, Paris, France.
Soller, M. and J.S. Beckmann, 1983. Genetic polymorphism in varietal identification 
and genetic improvement. Theoretical and Applied Genetics. 67:25-33.
Sprague, G.F., and L.A. Tatum, 1942. General versus specific combing ability in single 
crosses of corn. Journal of the American Society of Agronomy 34:923-932.
Stam, P. 1993. Construction of integrated genetic linkage maps by means of a new 
computer package: JoinMap. Plant Journal 3: 739-744.
Stanley, J., T. Frischmuth and S. Ellwood, 1990. Defective DNA amelioration symptoms 
of geminivirus infection in transgenic plants. Proceedings National Academy of 
Science (USA) 87:6291-6295.
Staskawicz, B.J, F.M. Ausbel, B.J. Barker, J.G. Ellis and J.D.G. Jones, 1995. Molecular 
genetics of plant disease resistance. Science 268:661-667.
Staub, J.E, J.C. Serquen and M. Gupta, 1996. Genetic markers, mapconstruction and 
their application in plant breeding. HortScience 31(5):729-740.
Storey, H.H. and R.F.W. Nichols, 1938. Studies on the mosaic disease of cassava. 
Annals of Applied Biology, 25:790-806.
Suiter, .A., J.F. Wendell, and J.S.S. Case. 1983. Linkage-1. A PASCAL computer 
program for the detection and analysis of genetic linkage. Journal of Heredity 
74: 203-204.
Tanksley, S.D., 1993. Mapping polygenes. Annual Review of Genetics 27:205-233.
Tanksley, S.D., R. Bernatzsky, N.L. Lapitan, and J.P. Prince, 1988. Conservation of 
gene repertoire but not gene order in pepper and tomato. Proceedings of 
National Academy of Science USA 85: 6419-6423.
Tanksley, S.D., N.D. Young., A.H Paterson and M.W. Bonierbale, 1989. RFLP mapping 
in plant breeding: New tools for an old science. Bio/Technology 7:257-264.
Terry, E.R. and S.K. Hahn, 1980. The effect of cassava mosaic disease on growth and 
yield of a local and an improved variety of cassava. Tropical Pest Management 
26:34-37.
196
Theberge, R.L., F.E Caveness, K.M. Lema, and J.S. Yaninek, 1985. Common African 
pests and diseases of cassava, yam, sweetpotato and cocoyam. IITA, Ibadan, 
Nigeria. 108pg.
Thankappan, M. and C.l. Chacko, 1976. Effect of cassava mosaic on the different plant 
parts and tuber yield in cassava. Journal of Root Crops 2: 45-47.
Thoday, J.M. 1961. Location of polygenes. Nature 191:368-370.
Thomas ,C.M, D.A. Jones, M. Parniske, K. Harrison, P.J. Balint-Kurt, K. Hatzixanthis 
and J.D.G. Jones, 1997. Characterisation of the tomato Cf-4 gene for resistance 
to Cladosporium fulvum identifies sequences that determine recognition 
specificity in Cf-4 and Cf-9. The Plant Cell 9:2209-2224.
Thomas, J. E. P.R. Massalski and B.D. Harrison, 1986. Production of monoclonal 
antibodies to African cassava mosaic virus and differences in their reactivities 
with other white-fly transmitted geminiviruses. Journal of General Virology 
67:2739-2748.
Thresh, J.M., D. Fargette and G.W. Otim-Nape, 1994a. Effect of African cassava 
mosaic geminivirus on yield of cassava. Tropical Science 34:26-42.
Thresh, J.M., L.D.C Fishpool, G.W. Otim-Nape and D. Fargette, 1994b. African cassava 
mosaic diseases: an underestimated and unsolved problem. Tropical Science 
34:3-14.
Thresh, J.M. and G.W. Otim-Nape, 1994. Strategies for controlling African cassava 
mosaic geminivirus. Advances in Disease Vector Research 10: 215-236.
Thresh, J.M., G.W. Otim-Nape and D. Fargette, 1998b. The components and 
deployment of resistance to cassava mosaic virus disease. Integrated Pest 
Management Reviews 3:209-224.
Thresh, J.M., G.W. Otim-Nape, J.P. Legg and D. Fargette, 1997. African cassava 
mosaic virus disease: The magnitude of the problem. African Journal of Root 
and Tuber Crops 2:13-18.
Thresh, J.M., G.W. Otim-Nape and D. Fargette, 1998a. The control of African cassava 
mosaic virus disease: phytosanitation and/or resistance? In: A. Hadidi, R.K. 
Khetarpal and H. Koganezawa (edsj, Plant Virus Disease Control, pg 670-677, 
APS Press. St. Paul. Minnesota, USA.
Umemura, Y. and K. Kawano, 1983. Field assessment and inheritance of resistance to 
cassava bacterial blight. Crop Science 23:1127-1132.
Vogel, J.M., A. Rafalski, W. Powell, M. Morgante, C. Andre. M. Hanafey, and S.V. 
Tingey. 1996. Application of genetic diagnostics to plant genome analysis and 
plant breeding. HortScience 31:1106-1107.
Vos, P., R. Hogers, M. Bleeker, M. Reijans, T. Van de Lee, M. Hornes, A. Frijters, J. 
Pot, J. Peleman, M. Kuiper and M. Zabeau, 1995. AFLP: a new technique for 
DNA fingerprinting. Nucleic Acid Research 21:4407-4414
Vuylsteike, D., 1981. Onderzoek op enkele agronomiche en fysiologische aspekten van 
cassava en zoete bataat. MSc Thesis, Katholiete Universiteit, Leuven, Belgium..
197
Wang, G.L, D.J., Mackill, J.M., Bonman, S.R., Couch, M.C., Champoux and R.J. 
Nelson. 1994. RFLP mapping of genes conferring complete and partial 
resistance to blast in a durably resistance rice cultivar. Genetics 136:1421-1434.
Wang, Y, R.L. Nelson and Y, Hui. 1998. Genetic analysis of resistance to soybean 
mosaic virus in four soybean cultivars from China. Crop Science 38:922-925.
Wang, G.L, D.L. Ruan, W.Y. Song, S. Sideris, L. Chen, L.Y. Pi, S. Zhang, Z. Zhang, C. 
Fauquet, B.S. Gaut, M.C. Whalen and P.C, Ronald. 1998. Xa21D encodes a 
receptor-like molecule with a leucine-rich repeat domain that determines race 
specific recognition and is subject to adaptive evolution. The Plant Cell 10:765- 
769.
Wanyera N.W.M., 1993. Phylogenetic relationships among cultivated cassava (Manihot 
esculenta Crantz) and two wild Manihot species. PhD Thesis University of 
Ibadan, Ibadan Nigeria 193 pg.
Welsh, J. and M. McClelland, 1990. Fingerprinting genomes using PCR with arbitrary 
primers. Nucleic Acid Research 18:7213-7218.
West, C.E., F. Pepping and C.R. Temalilwa, 1988. The composition of foods commonly 
eaten in East Africa. Wageningen Agriculture University. The Netherlands.
Wiiliams, C.G., 1998. QTL mapping in outbreed pedigrees. In; Paterson A.H (Ed). 
Molecular dissection of complex traits CRC Press CLL, pg 81-94, USA.
Wiliams, J.G.K., A.R. Kubelik, K.I. Livak, J.A. Rafalski and S.V. Tingey, 1990. DNA 
polymorphism amplified by arbitrary primers are useful as genetic markers. 
Nucleic Acid Research 18:6531-6535.
Wong, H.L. H.H. Yeoh, S.W. Lim, 1999. Customisation of AFLP analysis for cassava 
varietal identification. Phytochemistry 50:919-924
Wricke, G. and W.E. Weber, 1986. Quantitative genetics and selection in plant 
breeding. Walter de Gruyter, New York, USA. 406pg.
Wright,. S., 1968. Evolution and Genetics of Populations, vol. 1. Genetic and Biometric 
Foundations. University of Chicago Press, Chicago, USA.
Wu, K.K, W. Burnquist, M.E. Sorrels, T.L. Tew, P.H. Moore, S.D. Tanksley, 1992. The 
detection and estimation of linkage in polyploids using single dose restriction 
fragments. Theoretical and Applied Genetics 81:471 -476.
Xiao, J., J. Li, L. Yuan and S.D. Tanksley, 1995. Dominance is the major genetic basis 
of heterosis in rice as revealed by QTL analysis using molecular markers. 
Genetics 140:745-754.
Xu, Y., S. R. McCouch, and S. Zongtan 1998. Transgressive Segregation of Tiller Angle 
in Rice Caused by Complementary Gene Action. Crop Sciences 38:12-19.
Yaninek, J.S., G.J. de Moraes, and R.H. Markham, 1989. Handbook on the cassava 
green mite Mononychellus tanajoa in Africa. International Institute of Tropical 
Agriculture, Ibadan, Nigeria. 210 pg.
198
Y oung , N .D ., 1996. Q TL  m app ing  and  q u a n tita tiv e  d isease  re s is ta n ce  in p lan ts . Annual. 
Review of Phytopathology 34:479-501.
Y oung , N .D ., D. Z am ir, M .W . G ana , and  S .D . T an ks le y , 1988. U se  o f nea r is o g e n ic  
lines  and  s im u ltaneous  p rob ing  to  id en tify  D N A  m a rke rs  t ig h tly  lin ked  to  th e  T m - 
2a  gene  in tom a to . Genetics 120:579-585
Yu, S .B ., J .X  Li, Y .E . Tan, Y .J. G ao , X .H . Li, O .F . Z ha n g  a nd  M .A  S agha i-M a roo f, 1998. 
Im po rta nce  o f  e p is ta s is  as  th e  g e n e tic  b a s is  o f h e te ro s is  in an e lite  rice  hyb rid . 
Proceedings National Academy of Science (USA) 94:9226-9231.
Zeng , Z., 1992. C o rre c tin g  th e  b ia s  o f W r ig h t ’s e s tim a te  o f the  n u m b e r o f g enes  
a ffe c ting  a q u a n tita tive  cha ra c te r: A  fu rth e r  im p ro ved  m ethod . Genetics 131:987- 
1001 .
Zeng , Z .B ., D. H ou le , and  C .C . C o cke rh am , 1990. H ow  in fo rm a tive  is W r ig h t ’s e s tim a to r 
o f th e  n um be r o f  g enes  a ffe c ting  a q u a n tita tiv e  c h a ra c te r?  Genetics 126:235- 
247.
Zhang , G ., E .R . A nge ls , M .L.P . A b e ne s , G .S . K lush  and  N. H uang , 1996. R A P D  and  
R F LP  m app in g  o f  th e  bac te ria  b ligh t res is ta n ce  gene  xa -1 3  in rice. Theoretical 
and Applied Genetics 93:65-70.
Zhang , Y. and  M .S . Kang , 1997. D ia lle l-S A S : A  SA S  p rog ram  fo r  G r iff in g ’s  d ia lle l 
ana lyses . Agronomy Journal 89:176-182.
Zhou , X ., Y . Liu, L. C a lve rt, C . M unoz , G .W . O tim -N ape , D .J. R ob in son  and  B.D . 
H a rrison , 1997. E v idence  th a t D N A -A  o f  a g em in iv iru s  a sso c ia ted  w ith  se ve re  
ca ssa va  m osa ic  d ise a se  in U ganda  has  a risen  by in te r-sp e c ific  re com b ina tio n . 
Journal of General Virology 78:2101-2111.
199
Appendix I
Appendix la The physical and chemical properties of soil on which genetic experiments were planted
Soil parameter Mokwa 1998
Site 
Ibadan 1998 Ibadan 1999
Part sand (%) 76 89 82
Size silt (%) 9 5 13
Clay (%) 15 6 5
pH (H20  1:1) 5.4 6.1 5.6
Organic C (%) 0.7 0.96 1.69
Total N (%) 0.032 0.086 0.104
C/N ratio 8.4 11 16
Bray-1 P (mg/Kg) 0.4 5.2 18.2
Exchangeable Ca (cmol (+)/Kg) 1 1.2 3.5
Exchangeable Mg (cmol (+)/Kg) 0.3 0.3 0.7
Exchangeable K (cmol (+)/Kg) 0.03 0.2 0.4
Exchangeable Na (cmol (+)/Kg) 0.1 0.1 0.3
Exchangeable Mn (cmol (+)/Kg) 0.02 0.1 0.2
Exchangeable Acid (cmol (+)/Kg) 0.3 0.3 0.5
Effective cation exchange capacity (ECEC) 1.7 2.1 5.4
Cu (ppm) 0.5 0.8 3.1
Zn (ppm) Trace 3 19.3
Mn (ppm) 30.4 37.1 89.6
Fe (ppm) 2.9 10.6 16.8
200
Appendix lb Rainfall and temperature data for the Mokwa 1998/1999, Ibadan 1998/1999 and Ibadan 1999/2000 growing seasons*
Mokwa 1998/1999 Ibadan1998/1999 Ibadan 1999/2000
Total Temperature °C Total Temperature °C Total Temperature °C
Month Rainfall No Rainy Min Max Mean Rainfall No Rainy Min Max Mean Rainfall No Rainy Min Max Mean
(mm) Days (mm) Days (mm) Days
JUN 376.6 11 22.4 32.4 27.4 158.1 17 22.9 31 26.9 255.8 20 22.4 30.6 26.5
JUL 154.5 6 22.8 31.3 27.05 71.9 17 22.8 28.3 25.5 267 19 22.6 29.5 26.1
AUG 121 5 28.3 30.3 29.3 50.8 14 22.2 28.1 25.1 99.3 12 22.3 28.7 25.5
SEP 224.8 6 21.4 31.4 26.4 143.6 16 23 29 26 180.4 17 22.2 28.9 25.6
OCT 153.3 9 21.1 33 27.05 178.7 17 23.7 30.6 27.2 309.2 18 22.3 30.1 26.2
NOV 0 0 16 35.1 25.55 31.7 5 24.3 33.2 28.7 36.9 3 22.8 32.3 27.5
DEC 0 0 16.9 37 26.95 11.4 5 21.7 32.4 27 0 0 10.6 17.2 13.9
JAN 0 0 17.3 34.8 26.05 0 0 22.5 33.4 28 11.7 2 22.1 33.2 27.6
FEB 0 0 21.1 37.5 29.3 86.3 7 22.5 34.3 28.4 0 0 19.7 34.7 27.2
MAR 0 0 23.9 38.6 31.25 105.5 11 23.2 32.7 28 96.5 5 22.4 35.8 29.1
APR 38.5 3 25.1 38.6 31.85 176.8 14 23.4 32.4 27.1 123.5 10 22.9 32.8 27.9
MAY 139 8 23.2 34.2 28.7 130.8 18 22.9 32.3 27.6 87.3 7 22.6 32 27.3
JUN 148.3 6 23.7 33.6 28.65 255.8 20 22.4 30.6 26.5 163.9 11 21.7 30.1 25.9
JUL 215.5 6 23.5 31.6 27.55 267 19 22.6 29.5 26.1 231.6 12 21.4 28.4 24.9
* The data were made available by the Agroclim ato logy Unit o f the In ternational Institute o f T ropica l Agricu lture, Ibadan, Nigeria.
201
Appendix Ila Pearson Correlation Coefficients for CMD severity responses in genetic material across environments
MST6WK MCMD6WK MST12WK MCMD12WK MST20WK MCMD20WK ST6WKI CMD6WKI ST12WKI CMD12WKI ST20WKI CMD20W
MST6WK 0.923** 0.866** 0.865** 0.670** 0.644’ * 0.849** 0.666** 0.761** 0.697** 0.635** 0.509**
MCMD6WK 0.923** 0.816** 0.879** 0.607** 0.588** 0.869** 0.823** 0.737** 0.755** 0.618** 0.486**
MST12WK 0.866** 0.816** 0.931** 0.803** 0.840** 0.699** 0.534** 0.894** 0.757** 0.740** 0.702**
MCMD12WK 0.865** 0.880** 0.931** 0.702** 0.712** 0.744** 0.637** 0.876 0.865** 0.709** 0.622**
MST20WK 0.670** 0.607** 0.803** 0.702** 0.868** 0.532** 0.391** 0.711** 0.610** 0.912 0.739**
MCMD20WK 0.644** 0.589** 0.840** 0.713** 0.868** 0.530** 0.403** 0.793** 0.664** 0.807** 0.8942**
ST6WKI 0.848** 0.869** 0.699** 0.744** 0.532** 0.530** 0.838** 0.774** 0.770** 0.614** 0.490“
CMD6WKI 0.666** 0.822** 0.534** 0.636** 0.390** 0.403** 0.838** 0.635** 0.704** 0.520** 0.397**
ST12WKI 0.761** 0.737** 0.894** 0.875** 0.711** 0.793** 0.774** 0.635** 0.897** 0.767** 0.760**
CMD12WKI 0.697** 0.755** 0.757** 0.865** 0.609** 0.664** 0.770** 0.704** 0.897** 0.666** 0.659**
ST20WKI 0.635** 0.618** 0.739** 0.708** 0.912** 0.807** 0.614** 0.520 0.767** 0.666** 0.828
CMD20WKI 0.509** 0.485** 0.702** 0.621** 0.738** 0.894** 0.489** 0.397** 0.760** 0.659** 0.828**
MST6WK Mean shoot-tip CMD severity at 6 WAP
MST12WK Mean shoot-tip CMD severity at 12WAP
MST20WK Mean shoot-tip CMD severity at 20 WAP
MCMDSWK Mean whole plant CMD severity at 6 WAP
MCMD12WK Mean whole plant CMD severity at 12WAP
MCMD20WK Mean whole plant CMD severity at 20 WAP
ST6WKI
ST12WKI
ST20WKI
CMD6WKI
CMD12WKI
CMD20WKI
Mean shoot-tip CMD incidence at 6 WAP 
Mean shoot-tip CMD incidence at 12WAP 
Mean shoot-tip CMD incidence at 20 WAP 
Mean whole plant CMD incidence at 6 WAP 
Mean whole plant CMD incidence at 12WAP 
Mean whole plant CMD incidence at 20 WAP
202
Appendix lib Pearson Correlation Coefficients for CMD severity responses in genetic material in Mokwa 1998
MST6WK MCMD6WK MST12WK MCMD12WK ST6WKI CMD6WKI ST12WKI CMD12WKI
MST6WK 0.891** 0.726** 0.760** 0.937** 0.695** 0.771** 0.703**
MCMD6WK 0.891** 0.683** 0.780** 0.890** 0.901** 0.785** 0.786
MST12WK 0.726** 0.683** 0.926** 0.571** 0.387** 0.952** 0.811**
MCMD12WK 0.760** 0.780** 0.926** 0.612** 0.500** 0.931** 0.937**
ST6WKI 0.937** 0.890** 0.571** 0.612** 0.829** 0.732** 0.686**
CMD6WKI 0.695** 0.901** 0.387** 0.500** 0.829** 0.631** 0.644**
ST12WKI 0.771** 0.785** 0.952** 0.931** 0.732** 0.631** 0.912**
CMD12WKI 0.703** 0.786** 0.811** 0.937** 0.686** 0.644** 0.912**
MST6WK Mean shoot-tip CMD severity at 6 WAP
MST12WK Mean shoot-tip CMD severity at 12WAP
MCMD6WK Mean whole plant CMD severity at 6 WAP
MCMD12WK Mean whole plant CMD severity at 12WAP
ST6WKI
ST12WKI
CMD6WKI
CMD12WKI
Mean shoot-tip CMD incidence at 6 WAP 
Mean shoot-tip CMD incidence at 12WAP 
Mean whole plant CMD incidence at 6 WAP 
Mean whole plant CMD incidence at 12WAP
203
Appendix lie Pearson Correlation Coefficients for CMD severity responses in genetic material in Ibadan 1998
MST6WK MCMD6WK MST12WK MCMD12WK MST20WK MCMD20WK ST6WKI CMD6WKI ST12WKI CMD12WKI ST20WKI CMD20WKI
MST6WK 0.944** 0.854** 0.815** 0.671** 0.644** 0.900** 0.736** 0.688** 0.593** 0.635** 0.509**
MCMD6WK 0.944** 0.775** 0.813** 0.607** 0.589** 0.911** 0.883** 0.655** 0.641** 0.618** 0.486**
MST12WK 0.854** 0.775** 0.898** 0.804** 0.840** 0.713** 0.522** 0.897** 0.743** 0.740** 0.702**
MCMD12WK 0.815** 0.813** 0.898** 0.702** 0.713** 0.700** 0.609** 0.870“ 0.842** 0.709** 0.622**
MST20WK 0.671** 0.607** 0.804** 0.702** 0.868** 0.532** 0.391** 0.711** 0.610** 0.912** 0.739**
MCMD20WK 0.644** 0.589** 0.840** 0.713** 0.868** 0.530** 0.403** 0.793** 0.664** 0.807** 0.894**
ST6WKI 0.900** 0.911** 0.713** 0.700** 0.532** 0.530** 0.914** 0.689** 0.612** 0.614** 0.490**
CMD6WKI 0.736** 0.883** 0.522** 0.609** 0.391** 0.403** 0.914** 0.564** 0.609** 0.520** 0.397**
ST12WKI 0.688** 0.655** 0.897** 0.870** 0.711** 0.793** 0.689** 0.564** 0.895** 0.767** 0.760**
CMD12WKI 0.593** 0.641** 0.743** 0.842** 0.610** 0.664** 0.612** 0.609** 0.895** 0.666** 0.659**
ST20WKI 0.635** 0.618** 0.740** 0.709** 0.912** 0.807** 0.614** 0.520** 0.767** 0.666** 0.828**
CMD20WKI 0.509** 0.486** 0.702** 0.622** 0.739** 0.894** 0.490** 0.397** 0.760** 0.659** 0.828**
MST6WK
MST12WK
MST20WK
MCMD6WK
MCMD12WK
MCMD20WK
204
Mean shoot-tip CMD severity at 6 WAP 
Mean shoot-tip CMD severity at 12WAP 
Mean shoot-tip CMD severity at 20 WAP 
Mean whole plant CMD severity at 6 WAP 
Mean whole plant CMD severity at 12WAP 
Mean whole plant CMD severity at 20 WAP
ST6WKI
ST12WKI
ST20WKI
CMD6WK1
CMD12WKI
CMD20WKI
Mean shoot-tip CMD incidence at 6 WAP 
Mean shoot-tip CMD incidence at 12WAP 
Mean shoot-tip CMD incidence at 20 WAP 
Mean whole plant CMD incidence at 6 WAP 
Mean whole plant CMD incidence at 12WAP 
Mean whole plant CMD incidence at 20 WAP
Appendix lid Pearson Correlation Coefficients for CMD severity responses in genetic material in Ibadan 1999
MST6WK MCMD6WK MST12WK MCMD12WK ST6WKI CMD6WKI ST12WKI CMD12WKI
MST6WK 0.910** 0.836** 0.829** 0.694** 0.501** 0.591** 0.473**
MCMD6WK 0.910** 0.834** 0.878** 0.680** 0.547** 0.561** 0.549**
MST12WK 0.836** 0.834** 0.928** 0.613** 0.535** 0.785** 0.557**
MCMD12WK 0.829** 0.878** 0.928** 0.609** 0.557** 0.747** 0.675**
ST6WKI 0.694** 0.680** 0.613** 0.609** 0.658** 0.724** 0.710
CMD6WKI 0.501** 0.547** 0.535** 0.557** 0.658** 0.569** 0.641**
ST12WKI 0.591** 0.561** 0.785** 0.747** 0.724** 0.568** 0.785**
CMD12WKI 0.473** 0.549** 0.557** 0.675** 0.710** 0.641** 0.785**
MST6WK
MST12WK
MCMD6WK
MCMD12WK
205
Mean shoot-tip CMD severity at 6 WAP ST6WKI Mean shoot-tip CMD incidence at 6 WAP
Mean shoot-tip CMD severity at 12WAP ST12WKI Mean shoot-tip CMD incidence at 12WAP
Mean whole plant CMD severity at 6 WAP CMD6WKI Mean whole plant CMD incidence at 6 WAP
Mean whole plant CMD severity at 12WAP CMD12WKI Mean whole plant CMD incidence at 12WAP
Appendix III
Stock solutions and Buffers.
EDTA (0.5 M) pH 8.0
EDTA 186.1 g
NAOH 18g
Deionised water 1 L
Adjust pH to 8.2 Autoclave
TE (Tris EDTA)
10mMTris-HCI (pH 8.0) 0.1 mM EDTA 
Tris-HCI 1M pH 8.0 5 ml
EDTA 0.5 M 1 ml
Sterile deionised water up to 500
Tri Borate EDTA 10X
Tris base 108 g
EDTA0.5M 40 ml
Boric Acid 55 g
Deionised water up to 1 L
Tris-HC! 1M
Tris base 121.1 g
HCI pH 7.4 70 ml
 pH 7.6 60 ml
 pH 8.0 42 ml
Deionised water up to 1 L
TNE 10X
Tris base 121.1 g
EDTA-Na2 3.7 g
NaCI 58.4 g
Fluorometer Assay Buffer
Hoechst stock solution 10 pi
10X TNE buffer 10 ml
Sterile deionised water 90 ml
206
Gel electrophoresis reagents.
Appendix IV
Gel loading buffer
bromophenol blue, 
glycerol in water
Siliconising agents
Commercial automobile windscreen 
cleaner Rain-X 
Sigma cote or Gel slick
Gel Binding Solution
Bind silane 
Acetic Acid 95%
Ethanol 5%
6% Acrylamide
40% Acrylamide:bisacrylamide (19:1) 
Urea (7M)
10X TBE buffer
Deionised water up to 500ml and 
stored a 4°C.
To 75 ml of gel before use 
10% Ammonium persulphate 
Temed
Formamide dye
Bromophenol blue 
xylene cyanol in H20 
EDTA (0.5 M) 10 mM 
Formamide
Deionised water (sterile) up to 1000 pi
0.25 % 
40 %
1000 pi per plate 
600 pi per plate
3 pi 
1 ml
75 ml 
210 g 
50 ml
300 pi 
35 pi
0.25 %, 
0.25 % 
20 pi 
98 %
207
Silver staining reagents.
Append ix  V
Fix/stop solution
(10%  g la c ia l a c e tic  ac id ): 
G lac ia l a ce tic  a c id  
D e ion ised  w a te r
Staining solution
Silver Nitrate (AgN03) 
Formaldehyde 37%
Deionised water
Developing solution
Sodium carbonate (Na2C03) 
Deionised water 
Chill at 4°C. Just before use 
Formaldehyde 37%
Sodium thiosulfate (10mg/ml)
200 ml 
1,800 ml
2 g (1 packet)
3 ml (1 vial)
2 L
60 g (1 packet)
2 L
3 ml (1 vial) 
400 pi (vial)
208