Preliminary study on a yield-prediction model of maize (Zea mays L.) hybrid based on simple sequence repeat markers for breeding optimization by independent breeders in China

In China, the main breeding objective for maize ( Zea mays L.) is increasing the yield of single-cross hybrids. In this regard, developing yield-prediction models based on genetic markers for hybrids can enhance the probability of obtaining hybrid vigour in maize single-cross hybrids and reduce the cycle for germplasm development (inbred lines). In this study, we used simple sequence repeat markers to genotype 257 cross combinations from 97 commonly used maize inbred lines classified into four heterotic groups (Domestic Reid , P78599- type BSSS , Tangsipingtou , and Lvda Red Cob ). We calculated the Q values (the probability of each individual's genomic variation coming from each subpopulation) of each inbred line’s genetic components. We found that these reflected genetic distances between the parental inbred lines. The parental genetic difference was identified as a key factor influencing heterosis for yield performance of single-cross hybrids, and the interaction factors of Q values between the parents were found to be highly correlated with the accuracy of single-cross hybrid yield predictions. Moreover, we developed a yield-prediction model for maize single-cross hybrids based on our established equation: Y = 9480.2 − 2352.6R 1 R 2 − 1411.8R 1 L 2 + 94.1R 1 P 2 + 1148.0R 1 S 2 − 988.8L 1 R 2 − 1016.9L 1 L 2 − 655.7L 1 P 2 − 1175.4L 1 S 2 − 569.1P 1 R 2 + 371.6P 1 L 2 − 604.2P 1 P 2 + 1684.7P 1 S 2 + 733.1S 1 R 2 + 726.9S 1 L 2 + 924.2S 1 P 2 − 1678.1S 1 S 2 (the correlation coefficient r = 0.4778). Using this model for maize breeding, we achieved prediction accuracies of 66.7% and 76.9% for low and high-yielding single-cross combinations, thereby reducing the workload in field assessment experiments and improving breeding efficiency.