Genomic predictions identify superior soybean crosses

Yield improvement is critical for the commercial success of new cultivars. Choosing the pairs of lines to cross to achieve this is a daunting exercise, and the number of possible crosses vastly exceeds the number that can be handled. Although progeny performance can be predicted solely based on genetic makeup, the implementation of genomic selection awaits the demonstration that superior crosses can be efficiently identified using computer-simulated predictions.

In a Crop Science study, researchers report having predicted the performance of 60,000 soybean crosses, comparing the predicted performance of 101 of these to their success in producing improved lines for commercialization. The team found that all but two of the 22 crosses identified as superior by conventional selection had been predicted to display above-average yield. They also noted that 96.2% of all crosses predicted to produce below-average progeny had been eliminated during selection.

This study therefore indicates that genomic predictions of cross performance could help make cultivar development more efficient by either performing fewer crosses and achieving the same yield improvement with fewer resources or, alternatively, investing the same resources on a more promising set of crosses and thereby achieving greater improvement.