Genetics of Plant Architecture and Its Effect on Yield in Cassava (Manihot esculenta Crantz)

Abstract

Cassava (Manihot esculenta Crantz) is a significant starch source for various industrial applications. However, cassava's traditional cultivation and harvesting methods are labour intensive and inefficient, limiting the supply of fresh cassava roots for industrial starch production. Cassava cultivars with compact plant architecture and moderate plant height are required to increase the productivity and quality of fresh cassava roots through mechanical farming. Plant architecture-related parameters such as plant height, height at first branch, harvest index, branching level, stem diameter, branching angle, and lodging tolerance are crucial for crop yield and adaptability to mechanized cultivation. However, the genetics of cassava plant architecture still need to be better understood. The purpose of this study was to determine the genetic bases of the plant architecture, yield and productivity-related traits in cassava, especially starch content. A panel of 434 cassava clones developed at the International Institute of Tropical Agriculture, Nigeria, was genotyped and phenotyped for 17 plant architecture, yield and productivity-related traits at four locations across four growing seasons in Nigeria. which constitute five test environments. A genome-wide association study (GWAS) was conducted using the phenotypic data from a panel of 434 clones and 54,574 high-quality DArTseq-derived SNP markers evenly distributed across the cassava genome. Significant associations between 17 SNPs and eight plant architecture and yield component traits were identified through GWAS. Out of these 17 SNPs, 13 were linked to 18 putative candidate genes for seven traits including angle of branching (4), plant type (4), overall plant appearance (2), level of branching (4), harvest index (3), and number of harvested plants (1). These candidate genes exhibit various functions in relation to plant architecture, adaptation, yield, plant growth, development, stress response, and starch metabolism. One of the 18 putative candidate genes identified in this study is a novel gene (Manes.01G077900). This represents a significant contribution to knowledge. The effect of genotype by environment interaction on the stability of genotypes was evaluated using the AMMI (Additive Main Effects and Multiplicative Interaction), Finlay-Wilkinson, and GGE (Genotype Main Effects and Genotype by Environment Interaction) biplots while the test environments were delineated into mega environments using statistical models such as AMMI (Additive Main Effects and Multiplicative Interaction), and GGE (Genotype Main Effects and Genotype by Environment Interaction) biplots. The genotypic main effect showed sufficient variation that could be exploited for genetic gains, but the interaction between genotype and environment could complicate this. Two cassava accessions with erect plant type G424 (TMS18F1436P0049) and G83 (TMS18F1096P0013) were found to be adaptable and stable for fresh root yield and starch content, making them early selection candidates for these traits. The prospect of adopting the use of near infrared spectroscopy data obtained using the affordable SCiO sensor spectrometer in predicting starch and dry matter content from fresh cassava roots was evaluated. The results obtained from this research produced identical results to the prediction metrics that were reported in previous trials with respect to the consistency and accuracy of spectra (NIRS) data that were obtained using SCiO sensor spectrometer. The results validate the previous findings with desirable accuracy of prediction. A high throughput phenotyping procedure that enhances the rapid and cost-effective estimation of plant architecture and yield traits in cassava was developed using normalized difference vegetation index (NDVI) data obtained using an affordable handheld sensor manufactured by Trimble. Two models (linear regression and polynomial regression model) were used in developing phenotyping protocols for predicting yield and plant architecture traits in cassava using NDVI data. The polynomial regression model showed similar prediction accuracy for fresh root yield across two environments at 3, 6, and 9 months after planting. The linear regression model used NDVI data to predict yield or plant architecture traits directly. Based on prediction accuracy, there exist a significant disparity between NDVI data obtained from the two trial locations (Mokwa and Onne), this affirms the effect of the significant genotype by environment interaction on the performance of the cassava accessions across the test environments. The cassava accessions that combine the desirable plant architecture traits, high yield and more than 25% starch content would be selected for further evaluation at the advanced yield trial (AYT) stage and subsequent stages of the cassava improvement program.