Amenga‑Etego et al. Malar J (2021) 20:152 https://doi.org/10.1186/s12936‑021‑03693‑3 Malaria Journal RESEARCH Open Access Temporal evolution of sulfadoxine‑pyrimethamine resistance genotypes and genetic diversity in response to a decade of increased interventions against Plasmodium falciparum in northern Ghana Lucas N. Amenga‑Etego1* , Victor Asoala2, Godfred Agongo2, Christopher Jacob3, Sonia Goncalves3, Gordon A. Awandare1, Kirk A. Rockett3,4 and Dominic Kwiatkowski4,5 Abstract Background: Anti‑malarial drug resistance remains a key concern for the global fight against malaria. In Ghana sulfadoxine‑pyrimethamine (SP) is used for intermittent preventive treatment of malaria in pregnancy and combined with amodiaquine for Seasonal Malaria Chemoprevention (SMC) during the high malaria season. Thus, surveillance of molecular markers of SP resistance is important to guide decision‑making for these interventions in Ghana. Methods: A total of 4469 samples from uncomplicated malaria patients collected from 2009 to 2018 was submitted to the Wellcome Trust Sanger Institute, UK for DNA sequencing using MiSeq. Genotypes were successfully translated into haplotypes in 2694 and 846 mono infections respectively for pfdhfr and pfdhps genes and the combined pfhdfr/ pfdhps genes across all years. Results: At the pfdhfr locus, a consistently high (> 60%) prevalence of parasites carrying triple mutants (IRNI) were detected from 2009 to 2018. Two double mutant haplotypes (NRNI and ICNI) were found, with haplotype NRNI hav‑ ing a much higher prevalence (average 13.8%) than ICNI (average 3.2%) across all years. Six pfdhps haplotypes were detected. Of these, prevalence of five fluctuated in a downward trend over time from 2009 to 2018, except a pfdhps double mutant (AGKAA), which increased consistently from 2.5% in 2009 to 78.2% in 2018. Across both genes, pfdhfr/ pfdhps combined triple (NRNI + AAKAA) mutants were only detected in 2009, 2014, 2015 and 2018, prevalence of which fluctuated between 3.5 and 5.5%. The combined quadruple (IRNI + AAKAA) genotype increased in prevalence from 19.3% in 2009 to 87.5% in 2011 before fluctuating downwards to 19.6% in 2018 with an average prevalence of 37.4% within the nine years. Prevalence of parasites carrying the quintuple (IRNI + AGKAA or SGEAA) mutant haplo‑ types, which are highly refractory to SP increased over time from 14.0% in 2009 to 89.0% in 2016 before decreasing *Correspondence: lamenga‑etego@ug.edu.gh 1 West African Centre for Cell Biology of Infectious Pathogens, Department of Biochemistry, Cell and Molecular Biology, University of Ghana, Legon, Accra, Ghana Full list of author information is available at the end of the article © The Author(s) 2021. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creati veco mmons.o rg/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Amenga‑Etego et al. Malar J (2021) 20:152 Page 2 of 11 to 78.9 and 76.6% in 2017 and 2018 respectively. Though quintuple mutants are rising in prevalence in both malaria seasons, together these combined genotypes vary significantly within season but not between seasons. Conclusions: Despite high prevalence of pfdhfr triple mutants and combined pfdhfr/pfdhps quadruple and quintuple mutants in this setting SP may still be efficacious. These findings are significant as they highlight the need to continu‑ ously monitor SP resistance, particularly using deep targeted sequencing to ascertain changing resistance patterns. Keywords: Malaria, SP resistance, Genotypes, Haplotype, Pfdhfr, Pfdhps, Temporal Background and these haplotypes are often associated with higher lev- Malaria disproportionately leads to childhood morbidity els of resistance as compared to single mutant genotypes. and mortality in sub-Saharan Africa (sSA). An estimated Typically, the most prevalent multi-locus genotype found 24 million children were infected with Plasmodium fal- in areas of high SP resistance in sSA is the triple pfdhfr ciparum in 2018 in sSA with Ghana being one of top ten mutant genotype (pfdhfr-51I/59R/108N/I164) [16, 17]. A countries in Africa with the highest absolute increases in quadruple pfdhfr mutant first detected in southeast Asia cases of malaria in 2018 compared to the previous year in the late 1980s, includes an additional I164L substitu- [1]. Due to persistent high malaria transmission, the tion [18]. Parasites carrying the quadruple haplotype are Ghana National Malaria Control Programme (NMCP) associated with high SP treatment failure [19]. Thus far, prioritized the northern regions for high-impact inter- parasites carrying this allele have been only detected in ventions, such as indoor residual spraying (IRS), sea- parts of East Africa [16, 20], but not West Africa [11, 12]. sonal malaria chemoprevention (SMC) among children Parasites carrying the triple pfdhr mutant under 5 years old [2], increased coverage of long-lasting (108N/51I/59R/I164) alleles and additional pfdhps dou- insecticidal bed nets (LLINs) and intermittent preventive ble mutant (G437/540E) alleles have been associated with treatment in pregnancy (IPTp) [3, 4]. The World Health strong resistance to SP [21]. A further mutation at codon Organization (WHO) recommends use of SP plus amo- 581 of pfdhps results in a triple mutant allele (A437G/ diaquine (SP-AQ) for SMC in areas of high seasonal K540E/A581G), which in combination the pfdhfr triple malaria transmission in the Sahel sub-region of Africa mutant confers a high rate of SP treatment failure [22, [5]. Therefore, despite the withdrawal of SP in 2005 as 23]. The occurrence of the K540E mutation is a proxy a first-line anti-malarial in Ghana [6], it remains a key marker for a variant quintuple mutant genotype contain- component in interventions targeting vulnerable groups ing both the pfdhps double mutant (A437G/K540E) and such as pregnant women and young children. However, the pfdhfr triple mutant genotypes, which is also highly the success of targeted interventions such as, IPTp and correlated with SP treatment failure in children [9, 21, SMC, is largely dependent on the population prevalence 22]. of bifunctional dihydrofolate reductase-thymidylate syn- The objectives of the present study were to deter- thase (dhfr) and dihydropteroate synthetase (dhps) SP mine the temporal trends in prevalence of multi-locus resistance mutations [7–9]. SP resistance markers in P. falciparum isolates collected Resistance of P. falciparum to SP is caused by well doc- from clinical sources between 2009 and 2018, a period umented single nucleotide polymorphisms (SNPs) in the spanning major interventions for reducing the burden of pfdhfr and pfdhps genes that encode enzymes in the folate malaria and disrupting transmission including IRS, SMC metabolism pathway and are targeted by pyrimethamine and IPTp. This study also explores the seasonal risk of and sulfadoxine, respectively [10, 11]. A change from carrying SP resistant genotypes in northern Ghana. wild-type Ser108 to Asn108 (S108N) in pfdhfr is associ- ated with low level pyrimethamine tolerance both in vitro Methods and in vivo [10–12]. This basal S108N amino acid substi- Study site tution is characterized to have a tenfold increased risk of The study was conducted in the Kassena-Nankana East SP therapeutic failure [13]. The progressive accumulation Municipality and Kassena-Nankana West District, two of other mutations that result in altered amino acid sub- adjoining administrative areas in the Upper East Region stitutions including pfdhfr-C50R, pfdhfr-N51I, pfdhfr- of northern Ghana, here referred to as KNDs. The KNDs C59R, and pfdhfr-I164L that can result in higher levels of cover a total area of about 1675km2 and lie between lati- resistance to pyrimethamine and diminish the efficacy of tude 10.30′ and 11.10′ North and longitude 1.1′ West SP in vivo [14]. Parasites carrying multiple mutations in close to the Ghana-Burkina Faso border (Fig. 1). Average the pfdhfr gene have evolved independently in different rainfall is estimated at about 1300 mm, but this has been populations throughout the malaria endemic world [15], consistently dwindling in recent years [24]. The KNDs Amenga‑Etego et al. Malar J (2021) 20:152 Page 3 of 11 Fig. 1 Location of the Upper East Region in Ghana, The Kassena‑Nankana Districts in the Upper East Region and Kassena‑Nankana East and West Districts showing location of Health facilities from which samples were collected. CHPS; Community Health and Planning Services population is estimated at 160,000 inhabitants with a to May. The main malaria vector is Anopheles gambiae population density of 91.5 per square kilometre [25]. and the annual entomological inoculation rate (EIR) was The proportion of children under 5 years has been esti- estimated to vary between 1132 and 157 infective bites/ mated at about 11.0% and there are approximately 4000 person/year in the KNDs [27]. births per year. The KNDs population is under continu- ous demographic surveillance by the Navrongo Health Ethics statement and Socio-demographic Surveillance System. There is Scientific and ethical clearance was obtained from the one main hospital, the Navrongo War Memorial Hospital Navrongo Health Research Centre Institutional Review (NWMH) that serves as a referral unit for the population, Board (# NHRCIRB203). Informed consent was obtained nine satellite clinics and several additional community- from all individuals or their parents/guardians prior to based health planning and services compounds strategi- enrollment into each study. cally located to increase access to basic primary health care [26]. Malaria transmission is intense with seasonal Study procedure fluctuations largely dependent on the rainfall pattern All-age patients with fever (or history of fever within [24, 25]. The high transmission season coincides with the past 24hrs) reporting to the NWMH, and four clin- the short rainy season from July to November, while the ics in the KNDs were screened for malaria from Janu- low season occurs during the dry months of December ary to December of each year from 2009 to 2018, except Amenga‑Etego et al. Malar J (2021) 20:152 Page 4 of 11 2012 when no survey was conducted. Plasmodium fal- 2020. 07.2 3. 20159 624v1 and https:// www.m alar iagen. ciparum Histidine Rich Protein 2-based rapid diagnos- net/r esour ce/2 9). tic test (RDT)[CareStart™ malaria Pf (HRP2), Access Bio, NJ, USA] was used. Informed consent was docu- Within‑host diversity estimation mented for RDT positive patients and those willing to Within-host diversity of (complexity) infections was participate were enrolled. Four 50 µL dried blood spots determined using either the genome-wide Fws metric in (DBS) and malaria smears were prepared from a finger the 2009 to 2013 sequence data or THE REAL McCOIL to prick. Plasmodium falciparum density was determined estimate infection complexity in the 2014 to 2018 ampli- by microscopy and asexual parasites scored against 200 con data. The McCOIL’s categorical method, which uses white blood cells (WBCs) and smears were considered likelihood estimation to determine the number of dis- negative after examining 100 fields without detecting tinct parasite genome-wide haplotypes (strains) within asexual or sexual parasites. Parasite density per 200 each sample was implemented [30]. WBCs was converted to density per microlitre assum- ing 8000 WBCs. Haplotype reconstruction Parasite DNA was extracted from DBS samples using Owing to the high within host diversity, in the 2009 to the QIAamp DNA Midi kit (Qiagen, UK) as per the 2013 collections, genotypes were translated into haplo- manufacturer’s protocol. types for each isolate based on the major allele for pfd- hfr and pfdhps loci. The amplicon data from 2014 to 2018 however, enabled the genotyping of pure or mixed (het- Genotyping of Drug resistance markers erozygous) samples and the reconstruction of haplotypes Samples between the 2009 and 2013 studies were sub- for both pure and heterozygous samples, accounting for mitted to the Malaria Genomic Epidemiology Net- both major and minor clones in mixed infections. Only work (MalariaGEN; http:// www. malar iagen. net/) at samples with Fws > 0.95 and MOI = 1 were used for con- the Wellcome Sanger Institute, UK for whole genome struction of haplotypes across both genes to obtain pfd- sequencing as previously described [28]. Genotypes hfr/pfdhps combined genotypes. were assigned for variants with a read depth ≥ 5. Gen- otypes for the selected resistance genes were then Data and statistical analysis extracted, translated and annotated into known SP To determine the temporal trends of multi-locus SP resistance markers. For samples collected between 2014 resistance markers, a year-on-year fluctuation in SP and 2018, amplicon sequencing of SP resistance mark- resistance markers was examined. To further explore the ers was performed by MalariaGEN using an amplicon- effect of malaria interventions, the data was grouped into sequencing protocol (https:// www. medrx iv.o rg/ conte three windows to reflect periods of major interventions- nt/1 0.1 101/2 020. 07. 23. 201596 24v1 and https:// www. 2009–2011(pre-IRS and SMC), 2013–2015 (IRS) and malar iagen. net/ resou rce/ 29). In brief, locus-specific 2016–2018 (SMC). However, it is worth noting that IPTp multiplexed primers were designed using MPprimer and bednets were implemented across all study peri- software [29] with some modifications. Primers were ods. The following analysis approaches were explored: designed to generate amplicons each of between 190 a descriptive analysis of temporal trends in proportion and 250 bp and were assigned to one of three pools. A of isolates with SP resistant marker haplotypes, mixed two-step PCR protocol was used for each pool to first species infections and within-host diversity of infec- amplify the regions of interest in the parasite genome, tions. The χ2 test was used to compare categorical vari- followed by a second PCR to incorporate sequencing ables among groups. The non-parametric Kruskal–Wallis and unique sample-level and primer-pool multiplex- test was used for group comparisons as appropriate for ing adapters. Multiple samples were sequenced on the distribution. All statistical tests were two-tailed and a single MiSeq lane combining all 3 pools of the PCR statistical significance set at p < 0.05. Data were analysed amplicons. Sequenced samples were de-plexed using using the open source statistical software R version 3.4.1. the unique multiplexing adapter IDs and the individual sample-level CRAM files were aligned to a modified Results amplicon P. falciparum reference genome. Genotyping A total 4469 patients with uncomplicated malaria were of bi-allelic SNPs was undertaken using bcftools as well enrolled into this study in the years 2009–2018. Of these as custom scripts to determine genotypes which were patients, 2694 [(2009; n = 120); (2010; n = 184); (2011; translated and annotated into known drug resistance n = 55); (2013; n = 33); (2014; n = 323); (2015; n = 178); haplotypes (https:// www.m edrxi v. org/ conte nt/1 0. 1101/ (2016; n = 920); (2017; n = 560); (2018; n = 321)] were available for genetic analysis of pfdhfr and pfdhps loci. A A menga‑Etego et al. Malar J (2021) 20:152 Page 5 of 11 total of 846 samples [(2009; n = 57); (2010; n = 81); (2011; in Fig. 3. At the Pfdhfr locus, year-on-year prevalence of n = 24); (2013; n = 22); (2014; n = 73); (2015; n = 70); triple (IRNI) Pfdhfr mutants associated with pyrimeth- (2016; n = 237); (2017; n = 175); (2018; n = 107)] passed amine resistance is on a general decline though still high as mono-infections (Fws > 0.95 or MOI = 1) and were (> 60%) (Fig. 3a). Two double mutant haplotypes (NRNI included in genetic analysis across both genes (Table 1). and ICNI) were detected, with haplotype NRNI having a higher year-on-year prevalence compared to haplotype Clinical and demographic characteristics ICNI, except in 2013 when their prevalence estimates In the years 2009 to 2018, participants were aged between were similar (Fig.  3a). The annual prevalence of NRNI 3.7 to 21.5 years. Overall, there were more female patients fluctuated with time from 7.5 in 2009 to 25% in 2018 (56.3%) than males in the study period 2009 to 2018. The whilst prevalence of ICNI fluctuated over time from 3.3% results show a declining trend in parasite density from in 2009 to 2.2% in 2018. In general, seasonality analysis 2009 to 2018. Trends in monthly mean parasite densities showed no significant differences in the prevalence of from 2009 to 2018 show a similar unimodal peak during pfdhfr mutants between the wet (high) and dry (low) sea- the high malaria season (rainy months of July to Novem- sons (Fig. Additional file 6: S5). However, the pfdhfr tri- ber) (Fig.  2 and Additional file  1: Table  S1) (Additional ple mutants increased in prevalence (> 75%) from 2015 to file 2: Fig S1). 2016 (when SMC was introduced) during the dry season The median parasite density decreased fourfold from and fluctuated sharply to about 50% and 59% in 2017 and 38,280 (IQR; 62,320) in the 2009–2011 study period to 2018 respectively (Additional file 6: Fig. S5A). In the dry 10,600 (IQR; 52,780) in the 2013–2015 study period season comparing the double mutants (NRNI and ICNI), but increased to 17,680 (IQR; 49,680) in the 2016–2018 NRNI had about a 20-fold higher prevalence (≤ 20%) period with a significant decline in the distribution of than the ICNI mutant (< 1%) between 2011 to 2015 parasite density during peak transmission from the when prevalence of NRNI decreased and that of ICNI 2009–2011 study period and the two subsequent sur- increased slightly to narrow their prevalence difference in veys (2009–2011 vs 2013–2015 (P < 0.001) or 2016–2018, 2016, after which NRNI increased sharply to > 30% whilst P < 0.001) (Additional file 3: Fig. S2). Overall, the major- ICNI decreased gradually to < 2% in 2018. On the other ity (> 97.8%) of infections were P. falciparum mono infec- hand, during the wet season, whilst pfdhfr triple mutants tions with 2.2% being co-infections with Plasmodium declined similarly as observed in the dry season, albeit a malariae as determined from amplicon sequencing data. gradual decrease after 2016, the double mutant (NRNI) fluctuated more rapidly at between 3 to 20% from 2009 Within‑host diversity to 2016 when prevalence of the two mutants equalized to Within-host diversity in the earlier studies from 2009 about 2.5% (Additional file 6: Fig. S5B). Between 2016 to to 2013 ranged from 0.003 to 0.864 (mean 0.254 ± 0.247 2018, NRNI increased sharply to about 20% whilst ICNI SD), estimated by the inbreeding co-efficient Fws. Over- decreased to below 1%. The I164L amino acid substitu- all, 79.5% (198/249) had Fws ≤ 0.5 in the 2009–2013 tion associated with high SP treatment failure in some studies, indicating a high proportion of multi-genomic settings was not detected in the current study. infections with high potential for outcrossing. Interest- At the pfdhps locus, six different haplotypes were ingly, overall, in the aggregated data, infections were observed. Of these haplotypes 5 declined in preva- significantly more complex in the low malaria transmis- lence over time from between 12.5 and 56% to less than sion season (P = 0.022) (Additional file  4: Fig. S3A) but 12.5% with the exception of the pfdhps double mutant the year on year showed significantly higher diversity (C/S/AGKAA), which fluctuated between 2009 and 2011 only in 2011 (P < 0.001) (Additional file  5: Fig. S4A). In and increased consistently from about 9% in 2011 to 71% the latter studies from 2014 to 2018 infection complex- in 2016, decreased slightly to 69.5% and then increased to ity ranged from one to three clones, except in 2015 when about 78% in 2018 (Table 1 and Fig. 3b). The results from up to four clones were detected (Additional file  5: Fig. seasonality analysis showed a similar trend with the same S4B). A higher proportion of monoclonal infections was 5 haplotypes occurring in both seasons and showing a observed in the 2016–2018 studies in both seasons. general decline from 2009 to 2018, particularly after 2014 with haplotype ((A/S)GKAS) having higher prevalence Temporal trends in pfdhfr, pfdhps and the combined pfdhfr/ in dry season compared to the wet season. However, the pfdhps genotypes double mutant (C/S/AGKAA), increased in prevalence The temporal trends in the pfdhfr, pfdhps haplotypes sharply in both seasons rising from about 33% in 2014 to and pfdhfr/pfdhps combined haplotypes were examined a peak of about 84% during the dry (low) season in 2016 on a year-on-year analysis of fluctuations in the preva- and then decreasing to 69% in 2018. The double mutant lence of circulating haplotypes in these genes as shown Amenga‑Etego et al. Malar J (2021) 20:152 Page 6 of 11 Table 1 Temporal trends in prevalence of SP resistant haplotypes from 2009 to 2018 Gene Haplotypea Sampling year 2009 2010 2011 2013 2014 2015 2016 2017 2018 % (n/N) % (n/N) % (n/N) % (n/N) % (n/N) % (n/N) % (n/N) % (n/N) % (n/N) pfdhfr (N51I/C59R/S108N/I164L) NRNI 7.5 (9/120) 9.2 (17/184) 14.5 (8/55) 6.1 (2/33) 20.7 (67/323) 18.5 (33/178) 7.1 (65/920) 16.1 (90/560) 24.6 (79/321) ICNI 3.3 (4/120) 2.2 (4/184) 1.8 (1/55) 6.1 (2/33) 1.2 (4/323) 1.1 (2/178) 5.7 (52/920) 4.8 (27/560) 2.2 (7/321) IRNI 80 (96/120) 77.2 (142/184) 76.4 (42/55) 66.7 (23/33) 67.5 (218/323) 73.6 (131/178) 79.8 (734/920) 70.7 (396/560) 67.9 (218/321) NCSI 9.2 (11/120) 11.4 (21/184) 7.3 (4/55) 21.1 (7/33) 10.5 (34/323) 6.7 (12/178) 7.5 (69/920) 8.4 (47/560) 5.3 (17/321) pfdhps (S436A[C/F/Y]/A437G/K540E/A581G/A613S) (A/F/Y/S)AKAS 14.2 (17/120) 19.0 (35/184) 18.2 (10/55) 6.1 (2/33) 7.1 (23/323) 3.4 (6/178) 3.4 (31/920) 4.1 (23/560) 6.9 (18/261) (C/S/A)GKAA 2.5 (3/120) 33.7 (62/184) 9.1 (5/55) 39.4 (13/33) 55.4 (179/323) 64.0 (114/178) 71.2 (655/920) 69.5 (389/560) 78.2 (204/261) AAKAA 55.0 (66/119) 22.8 (42/184) 41.8 (23/55) 21.2 (7/33) 21.7 (70/323) 7.3 (13/178) 2.2 (20/920) 9.8 (55/560) 8.4 (22/261) (A/S)GKAS 2.5 (3/120) 3.8 (8/184) 3.6 (2/55) 6.1 (2/33) 2.5 (8/323) 9.6 (17/178) 4.6 (42/920) 8.0 (45/540) 4.2 (11/261) (S/A)GEAA 0 (0/120) 0 (0/184) 0 (0/55) 3.0 (1/33) 0 (0/323) 1.7 (3/178) 1.1 (10/920) 1.3 0 (7/560) (0/261) SAKAA 25.8 (31/120) 20.7 (38/184) 27.3 (15/55) 24.2 (8/33) 13.3 (43/323) 14.0 (25/178) 17.5 (161/920) 7.3 (41/560) 2.3 (6/261) pfdhfr/pfdhpsb Triple 3.5 (2/57) 0 (0/81) 0 (0/24) 0 (0/22) 5.5 (4/73) 1.4 (1/70) 1.3 (3/237) 0 2.8 (0/175) (3/107) Quadruple 43.9 (25/57) 49.4 (40/81) 87.5 (21/24) 59.1 (13/22) 45.2 (33/73) 25.7 (18/70) 10.1 (24/237) 21.1 (37/175) 20.6 (22/107) Quintuple 52.6 (30/57) 50.6 (41/81) 12.5 (3/24) 40.9 (9/22) 49.3 (36/73) 72.9 (51/70) 88.6 (210/237) 78.9 (138/175) 76.6 (82/107) a Haplotypes derived from only pure samples. Quintuple are detected in all three studies and prevalence is much higher when heterozygous samples are included b Triple = NRNI + AAKAA; quadruple = IRNI + SGKAA or IRNI + AAKAA; quintuple = IRNI + AGKAA or IRNI + SGEAA; sextuple = IRNI + SGEGA A menga‑Etego et al. Malar J (2021) 20:152 Page 7 of 11 mutations (IRNI + AGKAA or SGEAA), which are highly refractory to SP initially decreased from 2009 (53%) to 2011 (13%) and then increased rapidly over time to 77% in 2018 with a light decrease in prevalence between 2016 (89%) and 2017 (79%) (Fig. 3c). No sextuple mutants were detected from 2009 to 2018. The pfdhfr/pfdhps combined genotypes did not differ in prevalence between seasons (p > 0.05) but within season prevalence estimates were statistically significant (Wilcoxon, dry season; p < 0.001 & wet season; p = 0.006) (Additional file 8: Fig. S7A). Also, only the prevalence of quintuple haplotype was signifi- cantly different between pre-SMC (2009–2015) and post- Fig. 2 Monthly trends in mean P. falciparum parasite density from SMC (2016–2018) (Additional file 8: Fig. S7B). However, 2009 to 2018 the year-on-year dry season prevalence of the quadru- ple mutants increased sharply from 2009 (9.8%) to 2011 also rose sharply from 46% in 2014 to peak at 85% 2016 (65%) but rapidly decreased from 2013 (60%) to 2016 and decrease to 82% in 2018 (Additional file 7: Fig. S6B). (10%) before increasing slightly to 19.8% in 2017 and fall- Figure  3c shows the yearly fluctuations in the com- ing back to 10% in 2018 (Additional file 9: Fig. S8A). The bined pfdhfr/pfdhps genotypes. Of the three haplotypes wet season dynamics showed marked fluctuation in the observed (i.e. triple, quadruple and quintuple), the preva- prevalence of quadruple mutants from slightly over 10% lence of triple (NRNI + AAKAA) haplotype was gen- in 2009 to 22% in 2011, and about 18% in 2014 to 14% in erally low (0 to 6%) from 2009 to 2018. The quadruple 2015 before increasing from about 2.5 to 9% in 2017 and (IRNI + AAKAA) mutants increased in frequency from 2018 respectively. No quadruple mutants were detected 2009 (44%) to 2011 (88%) [pre-IRS] and decreased rapidly in 2013 and 2016 during the wet season (Fig. S8B). On from 2011 to about 10% in 2016 before increasing sharply the one hand, during dry season, quintuple mutant prev- to about 21% in 2017 and 2018 (Table  1 and Fig.  3c). alence fluctuated markedly from 10% (2009), 50% (2010) However, prevalence of parasites carrying quintuple to 12% (2011), then increased gradually through 22% (2013), 28% (2014) to peak at about 48% prevalence in Fig. 3 Temporal trends in haplotype prevalence in SP resistance genes from 2009 to 2018. Panel A: pfdhfr double (ICNI or NRNI),triple (IRNI) and sensitive (NCSI), Panel B: pfdhps triple ([A/F/Y] AKAS or [A/S]GKAS or [S/AGEAA), double (C/S/A/GKAA), single (AAKAA) and fully sensitive (SAKAA). Panel C: pfdhr/pfhps combined haplotypes as defined in Table 1. Mutant amino acid substitutions are shown in bold Amenga‑Etego et al. Malar J (2021) 20:152 Page 8 of 11 2015 and 2016 before fluctuating down to 32% in 2017 particularly during the high malaria season in these set- and 38% in 2018. However, during the wet season quintu- tings [27]. ple mutants initially found at 5% prevalence were unde- In the absence of an alternative drug to SP for IPTp and tected consecutively in 2010 and 2011 but re-emerged other intermittent preventive treatment programmes in 2013 at 18% prevalence and increased consistently to in infants such as SMC, monitoring of resistance mark- peak at 46% in 2017 before dropping to a prevalence of ers in endemic populations is crucial. In characterizing 39% in 2018 (Additional file 9: Fig. S8B). SP-associated resistance loci in the current study set- tings, a persistently high prevalence of the triple-mutant (51I + 59R + 108  N + I164) pfdhfr haplotype from 2009 Discussion to 2018 was found. This haplotype, observed at > 60% The current study utilized data from Illumina whole prevalence year-on-year from 2009 to 2018, has been genome sequencing and deep amplicon sequencing previously demonstrated to contribute significantly to to determine SNP haplotypes for SP resistance mark- SP treatment failure [22]. The fairly stable double mutant ers following a stringent SNP calling criteria. This ena- haplotypes (NRNI and ICNI) appear to differ in fitness bled reliable haplotype reconstruction in both single with parasites carrying the NRNI haplotypes being more and multi-genome infections by accounting for minor prevalent. No marked seasonality was found in distribu- clones and avoiding the ambiguity posed when using the tion of these haplotypes year-on-year but the prevalence standard methodology or individual genotyping assays of NRNI was higher in dry (low) season suggesting the on mixed infections as observed in this study [31]. A sig- acquisition of the haplotype alleles by the infection res- nificant decline in the monthly mean prevalence of para- ervoir during high transmission and subsequent seeding sitaemia was observed between the 2009–2011 studies of infections in the dry season. Well-designed commu- and the subsequent studies in 2013–2015 or 2016–2018 nity studies are required to ascertain the seasonality during the high malaria transmission season. The 2009 to of these haplotypes. Of note is the complete absence of 2011 studies were conducted prior to the deployment of the pfdhfr-164L allele, which has been reported in parts strategic interventions, such as IRS and SMC in 2014 and of East Africa [33], but thus far not in West Africa. The 2016, respectively. Therefore, the marked decline in prev- counterpart SP-resistance associated gene locus pfdhps alence of parasitaemia in the post IRS study period 2016– also had several mutant haplotypes that have been previ- 2018 may be due to the direct impact of this intervention ously characterized to confer high resistance phenotypes. aimed at disrupting transmission. The combined effects Six different haplotypes were found in this gene, with two of IRS at the beginning of the high transmission season (AGEAA and AGKAS) containing markers of high-grade and the deployment of SMC (i.e. 4 rounds) during high resistance with AGKAS increasing over time and within transmission may explain these observations. However, seasons in this setting. The haplotype AGEAA, which has the observed significantly increased mean parasitaemia the K540E mutation previously shown to cause increased post-SMC (2013–2015 vs 2016–2018) corroborates our SP resistance is observed at < 5% prevalence year-on-year. observed shift in malaria morbidity to higher age groups, The declining year-on-year fluctuations of pfdhfr and which could be attributed to mass treatment campaigns pfdhps haplotypes detected suggest a slow expansion such as SMC that protect young children from clinical of these haplotypes in this study settings. This may be malaria. This highlights unintended epidemiological con- driven by high recombination rates, particularly dur- sequences of such targeted mass treatment campaigns ing the high malaria season. The higher complexity of and suggest that older children may also benefit from infections observed in the current study attest to high these campaigns. The high within-host diversity observed transmission intensity in these settings, which is ripe for in the 2009 to 2013 studies with over 70% of infections frequent recombination resulting in breakdown of these having Fws < 0.5 is an indication of transmission levels in long haplotypes. the era in these settings that pre-date major interventions Also, temporal trends in prevalence of pfdhfr/pfdhps such as IRS aimed at transmission reduction. The pre- combined genotypes show a fluctuating trend of quintu- sent study highlights the impact of the combined effects ple and quadruple mutants respectively until 2016 when of IRS, long-lasting bed nets and SMC, in driving down SMC was deployed in our study communities. Since then, multi-clonal infections with about 40% of infections har- the yearly prevalence of both combined haplotypes began bouring two clones per year from 2014 to 2018, and less to fluctuate in opposing directions with a slight decrease than 5% of infections have 3–4 clones between 2014 and in quadruple mutant prevalence and a slight increase in 2018. This is consistent with a drop in annual EIR esti- quintuple prevalence. It is interesting that the distribu- mates from previous > 250 infective bites/person/year tion of these combined genotypes was not significant [32] to about 50 bites/person/year with wide variation between the wet and dry season but rather within each A menga‑Etego et al. Malar J (2021) 20:152 Page 9 of 11 season (Additional file  8: Fig. S7A). This may be attrib- SP intervention programmes in Ghana. This study, how- uted to differences in transmissibility within each sea- ever, is limited to the catchment area of the Navrongo son. Only quintuple mutants were significantly different War Memorial Hospital, which is only one out of the 10 between pre-and post-SMC periods, suggesting a pos- NMCP sentinel sites for monitoring anti-malarial drug sible additional selection pressure imposed by the SMC efficacy across three transmission zones in Ghana [37]. campaign. This is supported by the observed increased Therefore, there is the need to assess the prevalence of year-on-year prevalence of quintuple mutants in the wet these molecular markers across all 10 sentinel sites with season (Additional file 9: Fig. S8). The overall survival of ecological variance across Ghana using targeted sequenc- parasites carrying quintuple mutants all year round may ing methods to make the data more valuable to the be attributed to increased fitness in the presence of rou- National Malaria Control Programme. tine anti-malarials and from prophylaxis under SMC. Previous studies have suggested that childhood interven- tion campaigns such as SMC are likely to promote the Conclusion spread of resistance in high transmission settings than This study highlights high prevalence of parasites carry- other adult interventions, such as IPTp-SP [34]. The SMC ing pfdhfr triple mutants and rising prevalence of com- drugs are given at the beginning of the rainy season (high bined pfdhfr/pfdhps quintuple mutants in this setting. transmission) in four rounds from July to October. These Thus, indicating the need to continuously monitor para- trends provide support for targeted interventions such as site response to SP across all three transmission zones in SMC to take place during the high transmission season Ghana. in order to maximize the long acting effect of AQ part- ner drug in this intervention to kill these parasites that may have a higher fitness in the presence of SP. However, AbbreviationsWHO: World Health Organization; SP: Sulfadoxine‑pyrimethamine; RDT: Rapid sentinel data from 2015 to 2017 showed PCR-corrected diagnostic test; NMCP: National Malaria Control Programme; SMC: Seasonal AS-AQ treatment efficacy of 98.2% in these settings [35], Malaria Chemoprevention; IPTp: Intermittent Preventive Treatment in Preg‑ and there is also accumulating evidence of rising preva- nancy; LLINs: Long‑lasting Insecticidal Nets; IRS: Indoor Residual Spraying; EIR: Entomological Inoculation Rate; SNP: Single nucleotide polymorphism; PCR: lence of AQ resistance markers in this population (data Polymerase chain reaction; Pfdhfr: Plasmodium falciparum dihydrofolate reduc‑ to be described elsewhere). Previous studies show that tase; Pfdhps: Plasmodium falciparum dihydropteroate synthetase. IPTp-SP may remain efficacious in the phase of circulat- ing high grade resistance genotypes but additional muta- Supplementary Information tions, such as 581G, 540E and 613S see the benefits of The online version contains supplementary material available at https:// doi. IPTp-SP begin to decline. Therefore, these mutations may org/ 10.1 186/ s12936‑0 21‑ 03693‑3. have distinct impacts on different malaria interventions (eg SMC, IPTp among others) and the genotype–pheno- Additional file 1: Table S1. type relations may not be perfect. Other factors, such as Additional file 2: Figure S2. Mean monthly P. falciparum parasitaemia from 2009 to 2018. amplification of parasite GTP cyclohydrolase 1 (GCH1) gene that mediate de novo folate biosynthesis [36], host Additional file 3: Figure S1. Distribution of parasite density by interven‑tion period. 1; is 2009–2011 (pre‑IRS) study, 2; is 2013–2015 study (IRS) and immunity and metabolism may limit these genotype– 3; is 2016–2018 study (SMC). High coverage of long‑lasting insecticidal phenotype interactions. bed nets from 2013. Overall, barring any sampling bias due to the passive Additional file 4: Figure S3 Overall seasonal distribution of Plasmodium case detection approach in this study, these findings falciparum complexity of infections. Panel A: Genome‑wide Fws metric from sequenced data (2009–2013). Panel B: Complexity of infection scored signal a real threat to obtaining the full benefit of key using COIL for amplicon data (2014–2018) studies. Study period 1‑high SP-based interventions, such as IPTp and SMC in this season and 2‑low season. setting. However, it is worth noting that the prevalence Additional file 5: Figure S4 Annual Seasonal distribution of Plasmodium levels obtained in the current study fall below the WHO falciparum complexity of infections from 2009–2018. Panel A: Genome‑ wide Fws metric from sequenced data (2009–2013). 1‑low season and limits of > 50% prevalence of 540E to signal failure of SP 2‑high season. Panel B: Complexity of infection scored using COIL for interventions at a particular setting [22]. amplicon data (2014–2018) studies. Preventing malaria in vulnerable groups such as preg- Additional file 6: Figure S5 Seasonal trends of pfdhfr haplotypes from nant women and young children comes with enormous 2009 to 2018. Panel A: Low; dry season (low malaria transmission season); benefits, yet there is currently no effective alternative to Panel B: High; Wet season (High malaria transmission season). SP for use in interventions targeting these groups. Hence, Additional file 7: Figure S6 Seasonal trends of pfdhps haplotypes from 2009 to 2018. Panel A: Low; dry season (low malaria transmission season); the need to continue to monitor these resistance markers Panel B: High; Wet season (High malaria transmission season). Hap1‑ in the general population and target populations to pro- (A/F/Y/S)AKAS; Hap2 ‑(C/S/A)GKAA; Hap3 – AAKAA; Hap4 ‑(A/S)GKAS; vide evidence-based guidance for the implementation of Hap5‑(S/A)GEAA; Hap6 ‑SAKAA Amenga‑Etego et al. Malar J (2021) 20:152 Page 10 of 11 Received: 26 November 2020 Accepted: 6 March 2021 Additional file 8: Figure S7 Distribution of pfdhfr/pfdhps combined genotypes during high (wet) and low (dry) seasons and pre‑and post‑SMC periods. Additional file 9: Figure S8 Temporal trends of pfdhfr/pfdhps combined genotypes during high (wet) and low (dry) seasons and pre‑and post‑SMC References periods. 1. WHO. World malaria report. Geneva: World Health Organization; 2019. 2. Nonvignon J, Aryeetey GC, Issah S, Ansah P, Malm KL, Ofosu W, et al. Cost‑ effectiveness of seasonal malaria chemoprevention in upper west region Acknowledgements of Ghana. Malar J. 2016;15:367. The authors would like to thank the Chiefs and people of the Kassena‑ 3. Steketee RW, Nahlen BL, Parise ME, Menendez C. The burden of malaria in Nankana Districts in the Upper East Region of northern Ghana, and the staff of pregnancy in malaria‑endemic areas. Am J Trop Med Hyg. 2001;64:28–35. the Navrongo Health Research Centre. We are extremely grateful to our study 4. Menéndez C, Bardají A, Sigauque B, Sanz S, Aponte JJ, Mabunda S, et al. participants and their caregivers. Our gratitude goes to Mr Sebastian Nsoh Malaria prevention with IPTp during pregnancy reduces neonatal mortal‑ who assessed the malaria blood smears, Mr Stephen Aberinga for supervising ity. PLoS ONE. 2010;5:e9438. the field work over the years. The authors will also like to thank the staff of 5. WHO. Policy recommendation: Seasonal malaria chemoprevention (SMC) Wellcome Trust Sanger Institute Sample Management, Genotyping, Sequenc‑ for Plasmodium falciparum malaria control in highly seasonal transmission ing and Informatics teams for their support. areas of the Sahel sub‑region in Africa. Geneva: World Health Organiza‑ Genome sequencing was carried out at the Wellcome Sanger Institute tion; 2015. and we thank the staff of the Wellcome Sanger Institute Sample Logistics, 6. Duah NO, Matrevi SA, de Souza DK, Binnah DD, Tamakloe MM, Opoku VS, Sequencing, and Informatics facilities for their contribution. et al. Increased pfmdr1 gene copy number and the decline in pfcrt and pfmdr1 resistance alleles in Ghanaian Plasmodium falciparum isolates after Authors’ contributions the change of anti‑malarial drug treatment policy. Malar J. 2013;12:377. LA, KR and DK conceptualised and design the study. LA, VA, GA supervised the 7. Iriemenam NC, Shah M, Gatei W, van Eijk AM, Ayisi J, Kariuki S, et al. field data collection. LA analysed the data and drafted the manuscript, GA, SG, Temporal trends of sulphadoxine‑pyrimethamine (SP) drug‑resistance CJ, KR and DK edited the manuscript. All authors read and approved the final molecular markers in Plasmodium falciparum parasites from pregnant manuscript. women in western Kenya. Malar J. 2012;11:134. 8. Venkatesan M, Alifrangis M, Roper C, Plowe CV. Monitoring antifolate Funding resistance in intermittent preventive therapy for malaria. Trends Parasitol. This work was supported by MalariaGEN and the Parasites and Microbes 2013;29:497–504. program at the Wellcome Sanger Institute with funding from the Wellcome 9. Kublin JG, Dzinjalamala FK, Kamwendo DD, Malkin EM, Cortese JF, Mar‑ Trust (206194; 090770/Z/09/Z) and by the MRC Centre for Genomics and tino LM, et al. Molecular markers for failure of sulfadoxine‑pyrimethamine Global Health which is jointly funded by the Medical Research Council and the and chlorproguanil‑dapsone treatment of Plasmodium falciparum Department for International Development (DFID) (G0600718; M006212). The malaria. J Infect Dis. 2002;185:380–8. Resource Centre for Genomic Epidemiology of Malaria is supported by 10. Roper C, Pearce R, Bredenkamp B, Gumede J, Drakeley C, Mosha F, et al. Wellcome (090770; 204911). Wellcome also provides core awards to the Antifolate antimalarial resistance in southeast Africa: a population‑based Wellcome Centre for Human Genetics (203141) and the Wellcome Sanger analysis. Lancet. 2003;361:1174–81. Institute (206194). Wellcome Trust (DELGEME grant 107740/z/15/z, Aspiring 11. McCollum AM, Poe AC, Hamel M, Huber C, Zhou Z, Shi YP, et al. Antifolate leader award). resistance in Plasmodium falciparum: multiple origins and identification of novel dhfr alleles. J Infect Dis. 2006;194:189–97. Availability of the data 12. Hastings MD, Bates SJ, Blackstone EA, Monks SM, Mutabingwa TK, Sibley Raw data is available on European Nucleotide Archive (ENA). All data gener‑ CH. Highly pyrimethamine‑resistant alleles of dihydrofolate reductase in ated is presented within the manuscript and its supplementary files. isolates of Plasmodium falciparum from Tanzania. Trans R Soc Trop Med Hyg. 2002;96:674–6. Declarations 13. Lozovsky ER, Chookajorn T, Brown KM, Imwong M, Shaw PJ, Kamchon‑wongpaisan S, et al. Stepwise acquisition of pyrimethamine resistance in the malaria parasite. Proc Natl Acad Sci USA. 2009;106:12025–30. Ethics approval and consent to participate 14. Mita T, Ohashi J, Venkatesan M, Marma ASP, Nakamura M, Plowe CV, et al. The study was conducted in accordance with the principles and Declara‑ Ordered accumulation of mutations conferring resistance to sulfadoxine‑ tion of Helsinki (1964) as amended in 2015. Ethical approval was obtained pyrimethamine in the Plasmodium falciparum parasite. J Infect Dis. from the Institutional Review Board of the Navrongo Health Research Centre 2014;209:130–9. (NHRCIRB203). Informed consent was obtained from all adult participants 15. Roper C, Pearce R, Nair S, Sharp B, Nosten F, Anderson T. Intercontinental and assent/parental consent for children and minors obtained through their spread of pyrimethamine‑resistant malaria. Science. 2004;305:1124. parents or Caregivers. 16. Naidoo I, Roper C. Mapping ‘partially resistant’, ‘fully resistant’, and ‘super resistant’ malaria. Trends Parasitol. 2013;29:505–15. Consent for publication 17. Tinto H, Guekoun L, Zongo I, Guiguemdé RT, D’Alessandro U, Ouédraogo All authors have consented for publication. JB. Chloroquine‑resistance molecular markers (Pfcrt T76 and Pfmdr‑1 Y86) and amodiaquine resistance in Burkina Faso. Trop Med Int Health. Competing interests 2008;13:238–40. The authors declare that they have no competing interest. 18. Sirawaraporn W, Sathitkul T, Sirawaraporn R, Yuthavong Y, Santi DV. Antifolate‑resistant mutants of Plasmodium falciparum dihydrofolate Author details 1 reductase. Proc Natl Acad Sci USA. 1997;94:1124–9. West African Centre for Cell Biology of Infectious Pathogens, Department 19. Hyde JE. Antifolate resistance in Africa and the 164‑dollar question. Trans of Biochemistry, Cell and Molecular Biology, University of Ghana, Legon, 2 R Soc Trop Med Hyg. 2008;102:301–3.Accra, Ghana. Navrongo Health Research Centre, Ghana Health Service, P. O. 3 20. Maiga OM, Kayentao K, Traoré BT, Djimde A, Traoré B, Diallo M, et al. Supe‑Box 114, Navrongo, Ghana. Wellcome Centre for Human Genetics, University 4 riority of 3 over 2 doses of intermittent preventive treatment with sulf‑of Oxford, Headington, Oxford OX3 7BN, UK. Wellcome Sanger Institute, 5 adoxine‑pyrimethamine for the prevention of malaria during pregnancy Wellcome Genome Campus, Hinxton, Cambridgeshire, UK. Big Data Institute, in Mali: a randomized controlled trial. Clin Infect Dis. 2011;53:215–23. University of Oxford, Oxford, UK. 21. Omar SA, Adagu IS, Gump DW, Ndaru NP, Warhurst DC. Plasmodium falciparum in Kenya: high prevalence of drug‑resistance‑associated A menga‑Etego et al. Malar J (2021) 20:152 Page 11 of 11 polymorphisms in hospital admissions with severe malaria in an epi‑ 31. Auburn S, Campino S, Miotto O, Djimde AA, Zongo I, Manske M, et al. demic area. Ann Trop Med Parasitol. 2001;95:661–9. Characterization of within‑host Plasmodium falciparum diversity 22. Naidoo I, Roper C. Drug resistance maps to guide intermittent preventive using next‑generation sequence data. Snounou G, editor. PLoS One. treatment of malaria in African infants. Parasitology. 2011;138:1469–79. 2012;7:e32891. 23. Matondo SI, Temba GS, Kavishe AA, Kauki JS, Kalinga A, van Zwetselaar 32. Appawu M, Owusu‑Agyei S, Dadzie S, Asoala V, Anto F, Koram K, et al. M, et al. High levels of sulphadoxine‑pyrimethamine resistance Pfdhfr‑ Malaria transmission dynamics at a site in northern Ghana proposed for Pfdhps quintuple mutations: a cross sectional survey of six regions in testing malaria vaccines. Trop Med Int Health. 2004;9:164–70. Tanzania. Malar J. 2014;13:152. 33. Gesase S, Gosling RD, Hashim R, Ord R, Naidoo I, Madebe R, et al. High 24. Koram KA, Owusu‑Agyei S, Fryauff DJ, Anto F, Atuguba F, Hodgson A, resistance of Plasmodium falciparum to sulphadoxine/pyrimethamine in et al. Seasonal profiles of malaria infection, anaemia, and bed net use Northern Tanzania and the emergence of dhps resistance mutation at among age groups and communities in northern Ghana. Trop Med Int codon 581. PLoS ONE. 2009;4:e4569. Health. 2003;8:793–802. 34. O’Meara WP, Smith DL, McKenzie FE. Potential impact of intermittent 25. Oduro AR, Wak G, Azongo D, Debpuur C, Wontuo P, Kondayire F, et al. preventive treatment (IPT) on spread of drug‑resistant malaria. PLoS Med. Profile of the Navrongo Health and Demographic Surveillance System. Int 2006;3:e141. J Epidemiol. 2012;41:968–76. 35. Harrington WE, Mutabingwa TK, Muehlenbachs A, Sorensen B, Bolla 26. Dalaba MA, Akweongo P, Savadogo G, Saronga H, Williams J, Sauerborn MC, Fried M, et al. Competitive facilitation of drug‑resistant Plasmodium R, et al. Cost of maternal health services in selected primary care centres falciparum malaria parasites in pregnant women who receive preventive in Ghana: a step down allocation approach. BMC Health Serv Res. treatment. Proc Natl Acad Sci USA. 2009;106:9027–32. 2013;13:287. 36. Heinberg A, Kirkman L. The molecular basis of antifolate resistance in 27. Kasasa S, Asoala V, Gosoniu L, Anto F, Adjuik M, Tindana C, et al. Spatio‑ Plasmodium falciparum: looking beyond point mutations. Ann N Y Acad temporal malaria transmission patterns in Navrongo demographic Sci. 2015;1342:10–8. surveillance site, northern Ghana. Malar J. 2013;12:63. 37. Abuaku B, Duah‑Quashie NO, Quaye L, Matrevi SA, Quashie N, Gyasi A, 28. Manske M, Miotto O, Campino S, Auburn S, Almagro‑Garcia J, Maslen G, et al. Therapeutic efficacy of artesunate–amodiaquine and artemether– et al. Analysis of Plasmodium falciparum diversity in natural infections by lumefantrine combinations for uncomplicated malaria in 10 sentinel sites deep sequencing. Nature. 2012;487:375–9. across Ghana: 2015–2017. Malar J. 2019;18:206. 29. Shen Z, Qu W, Wang W, Lu Y, Wu Y, Li Z, et al. MPprimer: a program for reliable multiplex PCR primer design. BMC Bioinform. 2010;11:143. 30. Chang H‑H, Worby CJ, Yeka A, Nankabirwa J, Kamya MR, Staedke SG, et al. Publisher’s Note The real McCoil: a method for the concurrent estimation of the complex‑ Springer Nature remains neutral with regard to jurisdictional claims in pub‑ ity of infection and SNP allele frequency for malaria parasites. PLOS lished maps and institutional affiliations. Comput Biol. 2017;13:e1005348. Ready to submit your research ? Choose BMC and benefit from: • fast, convenient online submission • thorough peer review by experienced rese archers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations • maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions