In plant breeding, finding the best genotypes from a big group is like finding a needle in a haystack. This study looks at different ways to make that process easier and more accurate. It's all about making the best use of genomic selection, which is like a superpower for plant breeders.

The study tested various methods to see which one could find the top-performing plants the fastest. They used metrics like normalized discounted cumulative gain, Spearman's rank correlation, and Pearson's correlation to see how well each method worked.

For populations without strong subgroups, a method called MSPE Ridge was the winner. It's like a special tool that helps find the best plants even when they're all mixed together.

But when the population has clear subgroups, other methods shine. A method called CDmean (v2) and another one that maximizes overall genomic variation, similar to D-optimality, were the stars. These methods help find the best plants even when they're grouped in different ways.

When there are lots of candidates, a new method called GVaverage can help narrow down the options. After that, a heuristic-based method can find the best of the best.

The study also found that a new version of CDmean (v2) is just as good as the original but works much faster.

Picking the Best Plants: How to Train Your Genomic Selection

Actions