Related Experiment Video
Updated: Aug 6, 2026
07:10
Untargeted Liquid Chromatography-Mass Spectrometry-Based Metabolomics Analysis of Wheat Grain
Published on: March 13, 2020
Multi-omics prediction for yellow rust in bread and durum wheat through conventional and Ai-based frameworks
Paolo Vitale1, Jose Crossa2, Abelardo Montesinos-López3
1International Maize and Wheat Improvement Center (CIMMYT), Carretera México-Veracruz, Km 45, 5623, El Batan, Edo. de México, Mexico.
Scientific Reports
|July 17, 2026
Summary
Unmanned aerial vehicle (UAV)-based high-throughput phenotyping (HTP) significantly improves yellow rust (YR) severity prediction in wheat compared to genomics alone. Integrating both data types did not further enhance accuracy, highlighting HTP
Area of Science:
- Plant Pathology
- Agricultural Science
- Genomics and Phenomics
Background:
- Yellow rust (YR) poses a significant threat to global wheat production, causing substantial yield losses.
- Traditional visual assessment of YR severity is labor-intensive, time-consuming, and susceptible to human error.
- Accurate and efficient YR severity prediction is crucial for effective disease management and breeding strategies.
Purpose of the Study:
- To evaluate the predictability of yellow rust severity using genomic and phenomic data in wheat.
- To compare the predictive performance of different modeling approaches, including machine learning and deep learning.
- To assess the potential of unmanned aerial vehicle (UAV)-based high-throughput phenotyping (HTP) for YR severity prediction.
Main Methods:
- Two biparental wheat populations (bread and durum) were used for the study.
- Yellow rust severity was visually scored, and UAV-based multispectral HTP data were collected.
- Genomic data (SNP arrays) and phenomic data (spectral wavelengths, vegetation indices) were analyzed using various predictive models.
Main Results:
- HTP-derived data significantly improved YR severity prediction accuracy (PA) compared to genomic markers alone.
- Support vector regression (SVR) showed a marked increase in PA from 0.35 (markers only) to 0.87 (wavelengths only).
- Integrating genomic and phenomic data did not yield further improvements, with models plateauing using HTP data alone.
- Cross-crop prediction demonstrated robust generalization between bread and durum wheat, achieving PA up to 0.83.
- Best linear unbiased prediction (BLUP) and multilayer perception (MLP) models performed consistently well across scenarios.
Conclusions:
- UAV-based HTP offers a rapid, scalable, and accurate method for predicting YR severity in wheat.
- Phenomic data, particularly spectral information, are highly effective for predicting YR severity.
- The integration of HTP and AI-driven prediction pipelines holds strong potential for wheat breeding programs.
