Genomic prediction of yield components in soybean under Asian soybean rust pressure for early-generation selection

Predicting the yield components of advanced lines based on the genotyping of plants in the initial generations can help to reduce the segregating lines that are carried forward in soybean breeding programs, especially for developing cultivars with high yields under Asian soybean rust (ASR) pressure. This study aimed to predict the average soybean yield components in the F2:3 and F2:5 families, as well as the top 50% lines of each F2:5 family [F2:5(50%] under ASR pressure based on genotyping of the F2 generation from a cross of two elite breeding lines. These lines have indeterminate growth habit and belong to maturity group 6. Phenotypes included seed yield per plant, 50-seed weight, days to maturity, and plant height. The genomic prediction models were G-BLUP, principal component regression, Bagging, and Bayes-Cπ. Bagging and Bayes-Cπ most often showed the highest predictive ability. Phenotyping of F2:3 as opposed to F2 only increased the predictive abilities of models for 50-seed weight. Phenotyping of F2:5 and F2:5 (50%) resulted in predictive abilities greater than 0.50 for all the traits. Therefore, F2 genotyping and genomic selection enabled prediction of soybean yield components in populations that are genetically similar to the target population and allowed optimizing the mechanical and financial resources of breeding programs to develop soybean cultivars with higher yields under ASR pressure.