Phenomic selection enables nursery screening for field performance in maize and pea

The article "Phenomic selection enables nursery screening for field performance in maize and pea" explores the potential of phenomic selection as an alternative to genomic selection in plant breeding programs. Traditionally, breeding programs eliminate most selection candidates early on, before their productivity can be assessed in field trials.

Genomic selection addresses this issue by predicting field performance early on, but it requires large-scale genotyping, which is inaccessible for many species and breeding programs.

Phenomic selection presents a more accessible alternative by relying on inexpensive high-dimensional phenotyping data. However, previous applications have primarily focused on predicting genotypes already evaluated in field trials. This study investigates whether phenomic data acquired through unoccupied aerial vehicles (UAVs) from nursery-grown genotypes can predict their performance in field trials.

The researchers validated this approach in maize and pea, two species with differing biology and canopy structures, suggesting the potential transferability of the method to other species.

In challenging scenarios, the average accuracy for yield-related traits in maize was 0.36, while for pea, it was 0.40. These accuracies indicate that UAV-based phenomic selection could support substantial nursery culling while minimizing the risk of discarding superior genotypes. The simulations also suggest that this proof of concept demonstrates the feasibility of using UAV-based phenomic selection to predict the field performance of nursery-grown genotypes without the need for genotyping the selection candidates.