Year: 2026
Type: preprint
Abstract: Due to economical constraints, breeding programs eliminate most selection candidates at early stages, before their productivity can be evaluated in field trials. Genomic selection partially overcomes this limitation by enabling early prediction of field performance, but its implementation requires large-scale recurrent genotyping that remains inaccessible for many species and breeding programs. Phenomic selection offers a promising alternative by replacing genotyping with inexpensive high-dimensional phenotyping data. However, previous applications have focused on predicting genotypes already evaluated in field trials. We investigated whether phenomic data acquired by unoccupied aerial vehicles (UAVs) from nursery-grown genotypes can be used to predict their performance in field trials. We validated this nursery-screening approach in maize and pea, two species with contrasting biology and canopy structures, supporting its potential transferability to other species. In the most challenging scenarios, average accuracies for yield-related traits were between 0.36 for maize and 0.40 for pea. Simulations indicated that these accuracies can support substantial nursery culling while limiting the risk of discarding superior genotypes. This proof of concept demonstrates that UAV-based phenomic selection can predict the field performance of nursery-grown genotypes without genotyping the selection candidates.
Institutions: AgroParisTech, Université Paris-Saclay, Institut National de Recherche pour l'Agriculture, l'Alimentation et l'Environnement, Génétique Quantitative et Évolution Le Moulon, KWS Saat (Germany), Centre de Coopération Internationale en Recherche Agronomique pour le Développement, Université de Montpellier, Amélioration Génétique et Adaptation des Plantes méditerranéennes et tropicales
Language: English
Subfield: Genetics
Domain: Life Sciences
SDG: Zero hunger
Open Access status: green
OpenAlex ID: W7222634024