An open field phenomics resource for multimodal maize yield prediction across divergent environments
- 1Texas A and M University
Abstract
Temporal drone phenotyping captures crop development, but irregular flight schedules complicate comparisons across environments. We release curated imagery from 356 flights across 19 Genomes to Fields environments containing 1,180 maize (Zea mays L.) hybrids. To evaluate its utility, we integrated functional principal components of vegetation index and weather trajectories with genomic information. Combinined genomic and phenomic kernels improved yield prediction, reaching correlations up to r = 0.501 for held-out hybrids in environments represented in training and 0.408 when environments were also withheld. Accumulated growing degree days offered no consistent predictive advantage over days after planting, and weather contributed modest, task-dependent gains. A transformer neural process learned directly from irregular observations, serving as a novel application of neural process models in agriculture. Mapping vegetation index functional principal components identified recurrent quantitative trait loci on chromosomes 3 and 7. This resource and its reproducible analyses guide the use of temporal spectral data for crop prediction and genetic discovery.
Related
- Integrating Envirotyping and Phenomics for AI-Enabled Multi-Environment Genomic Prediction in Crop Breeding
- An integrated genome-wide resource reveals distinct replication environments of DNA breakage in human cancer cell lines
- RBApy: Extending resource allocation modeling to eukaryotes in complex environments
- Whole-genome sequencing reveals divergent and shared selection signatures of heat stress adaptation in indigenous Ethiopian zebu cattle from dry-hot and humid-hot environments
- Beyond prediction accuracy by integrating GWAASB into wheat stability assessment across environments in Saudi Arabia
BibTeX
@misc{desalvio2026open,
title = {{An open field phenomics resource for multimodal maize yield prediction across divergent environments}},
author = {A. J. DeSalvio and P. Mohseni and A. Adak and S. C. Murray and M. A. Arik and R. K. W. Wong and J. Jung and D. C. Lima and A. C. Aviles and E. S. Buckler and N. Duffield and J. Edwards and D. Ertl and S. Flint-Garcia and M. A. Gore and C. N. Hirsch and J. B. Holland and S. M. Kaeppler and J. Miller and C. Romay and J. C. Schnable and M. P. Singh and E. E. Sparks and A. Thompson and J. D. Washburn and T. Weldekidan and N. D. Winans and N. de Leon},
howpublished = {bioRxiv},
year = {2026},
month = sep,
doi = {10.64898/2026.09.17.752474},
url = {https://www.biorxiv.org/content/10.64898/2026.09.17.752474v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - An open field phenomics resource for multimodal maize yield prediction across divergent environments
AU - A. J. DeSalvio
AU - P. Mohseni
AU - A. Adak
AU - S. C. Murray
AU - M. A. Arik
AU - R. K. W. Wong
AU - J. Jung
AU - D. C. Lima
AU - A. C. Aviles
AU - E. S. Buckler
AU - N. Duffield
AU - J. Edwards
AU - D. Ertl
AU - S. Flint-Garcia
AU - M. A. Gore
AU - C. N. Hirsch
AU - J. B. Holland
AU - S. M. Kaeppler
AU - J. Miller
AU - C. Romay
AU - J. C. Schnable
AU - M. P. Singh
AU - E. E. Sparks
AU - A. Thompson
AU - J. D. Washburn
AU - T. Weldekidan
AU - N. D. Winans
AU - N. de Leon
PY - 2026
DA - 2026/09/17
PB - bioRxiv
DO - 10.64898/2026.09.17.752474
UR - https://www.biorxiv.org/content/10.64898/2026.09.17.752474v1
AB - Temporal drone phenotyping captures crop development, but irregular flight schedules complicate comparisons across environments. We release curated imagery from 356 flights across 19 Genomes to Fields environments containing 1,180 maize (Zea mays L.) hybrids. To evaluate its utility, we integrated functional principal components of vegetation index and weather trajectories with genomic information. Combinined genomic and phenomic kernels improved yield prediction, reaching correlations up to r = 0.501 for held-out hybrids in environments represented in training and 0.408 when environments were also withheld. Accumulated growing degree days offered no consistent predictive advantage over days after planting, and weather contributed modest, task-dependent gains. A transformer neural process learned directly from irregular observations, serving as a novel application of neural process models in agriculture. Mapping vegetation index functional principal components identified recurrent quantitative trait loci on chromosomes 3 and 7. This resource and its reproducible analyses guide the use of temporal spectral data for crop prediction and genetic discovery.
ER -