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

Abstract

Asian soybean rust (ASR) has the potential do severely reduce soybean productivity in numerous regions of the world. Developing productive cultivars under disease pressure in genetic breeding programs is limited by low precision in the early stages and the large number of genotypes evaluated in more advanced stages. Information about their subsequent generations can increase genetic gain through selection and reduce the number of segregating lines carried forward in programs. The objective of this work was to predict the average soybean yield components in the F2:3 and F2:5 families and the 50% best lines of each F2:5 family (F2:5(50%)). Genotyping of the F2 generation from the cross of two elite breeding lines and phenotyping of advanced generations (F2:3 and F2:5) under ASR pressure were used to obtain and validate genomic prediction models. Phenotypes included seed yield per plant, 50-seed weight, days to maturity, and plant height. The genomic prediction models used were G-BLUP, principal component regression, Bagging, and Bayes C[[EQUATION]]. Bagging and Bayes C[[EQUATION]] most often showed the highest predictive ability. Phenotyping of F2:3 as opposed to F2 only increased the predictive above of 50-seed weight, but not for the other traits. Phenotyping of F2:5 and F2:5(50%) allowed us to obtain predictive abilities greater than 0.50 for all traits. Using these models for early-generation selection in a target population will allow better adaptation of the logistical structure of breeding programs to develop productive soybean genotypes under ASR pressure and increase selection gain.