Improving Maize Grain Yield under Drought Stress and Non‐stress Environments in Sub‐Saharan Africa using Marker‐Assisted Recurrent Selection

ABSTRACT In marker‐assisted recurrent selection (MARS), a subset of molecular markers significantly associated with target traits of interest are used to predict the breeding value of individual plants, followed by rapid recombination and selfing. This study estimated genetic gains in grain yield (GY) using MARS in 10 biparental tropical maize ( Zea may L.) populations. In each population, 148 to 184 F 2:3 (defined as C 0 ) progenies were derived, crossed with a single‐cross tester, and evaluated under water‐stressed (WS) and well‐watered (WW) environments in sub‐Saharan Africa (SSA). The C 0 populations were genotyped with 190 to 225 single‐nucleotide polymorphism (SNP) markers. A selection index based on marker data and phenotypic data was used for selecting the best C 0 families for recombination. Individual plants from selected families were genotyped using 55 to 87 SNPs tagging specific quantitative trait loci (QTL), and the best individuals from each cycle were either intercrossed (to form C 1 ) or selfed (to form C 1 S 1 and C 1 S 2 ). A genetic gain study was conducted using test crosses of lines from the different cycles F 1 and founder parents. Test crosses, along with five commercial hybrid checks were evaluated under four WS and four WW environments. The overall gain for GY using MARS across the 10 populations was 105 kg ha −1 yr −1 under WW and 51 kg ha −1 yr −1 under WS. Across WW environments, GY of C 1 S 2 –derived hybrids were 8.7, 5.9, and 16.2% significantly greater than those of C 0 , founder parents, and commercial checks, respectively. Results demonstrate the potential of MARS for increasing genetic gain under both drought and optimum environments in SSA.