Prediction of late blight severity in a large panel of potato genotypes using low-altitude aerial images and machine learning methods
4 affiliations
- 1International Potato Center (CIP), Lima, Peru
- 2Universidad de Huánuco, Huánuco, Peru
- 3Pontificia Universidad Católica del Perú, Lima, Peru
- 4Programa Nacional de Investigación Científica y Estudios Avanzados - Prociencia-Concytec, Lima, Peru
Abstract
Potato (Solanum tuberosum L.) is a staple crop crucial to global food security, yet its production is severely threatened by late blight (LB), caused by Phytophthora infestans, one of the most destructive plant diseases worldwide. Breeding programs for LB resistance have traditionally relied on labor-intensive and subjective visual assessments, which limit scalability and consistency, particularly in early-generation trials. Unmanned aerial vehicle (UAV)-based remote sensing combined with machine learning (ML) offers a promising alternative for objective, high-throughput disease phenotyping. This study evaluated the potential of this technology to estimate LB severity at the genotype level across two large potato trials in Peru comprising 2,745 breeding clones and 492 genebank accessions. We compared vegetation index–based approaches with a ML framework that integrates K-means clustering and Kernel Ridge Regression (KRR). While NDVI correlated strongly with visually assessed LB severity, particularly at advanced stages, the multispectral KRR approach outperformed linear NDVI models by capturing nonlinear relationships between spectral response and disease progression. Best linear unbiased estimates (BLUEs) of LB severity from both models were strongly and consistently related to the area under the disease progress curve (AUDPC), particularly in later UAV acquisitions. Selection coincidence with AUDPC-based rankings was substantially higher at intermediate to advanced stages of disease progression, indicating that a single UAV inspection at those stages captures enough phenotypic variation to distinguish genotypes. UAV-based multispectral phenotyping integrated with ML, therefore offers a practical, scalable alternative to the time-consuming visual assessments traditionally used in potato breeding programs.
Related
- A comparative analysis of graph neural networks and classical machine learning for carbapenem-resistance prediction and genomic interpretation in Acinetobacter baumannii
- A novel LLM-based GPT-neo transformer with machine learning architecture for promoters prediction in Agrobacterium
- Machine Learning for Toxicity Prediction in Low-Sample Molecular Classes
- A Two-Stage ESM-Based Machine Learning Pipeline for Robust Hierarchically Evaluated Enzyme Function Prediction
- Deep learning-based genetic disease prediction with Monte Carlo (MC) dropout uncertainty estimation: a comparative study with classical machine learning baselines
BibTeX
@article{loayza2026prediction,
title = {{Prediction of late blight severity in a large panel of potato genotypes using low-altitude aerial images and machine learning methods}},
author = {Hildo Loayza and Johan Ninanya and Susan Palacios and Luis Silva and Fernando Pujaico Rivera and Javier Rinza and Manuel Gastelo and Mariela Aponte and Jan F. Kreuze and Hannele Lindqvist-Kreuze and Bettina Heider and Moctar Kante and David A. Ramírez},
journal = {PLOS One},
year = {2026},
month = nov,
doi = {10.1371/journal.pone.0347446},
url = {https://doi.org/10.1371/journal.pone.0347446},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Prediction of late blight severity in a large panel of potato genotypes using low-altitude aerial images and machine learning methods
AU - Hildo Loayza
AU - Johan Ninanya
AU - Susan Palacios
AU - Luis Silva
AU - Fernando Pujaico Rivera
AU - Javier Rinza
AU - Manuel Gastelo
AU - Mariela Aponte
AU - Jan F. Kreuze
AU - Hannele Lindqvist-Kreuze
AU - Bettina Heider
AU - Moctar Kante
AU - David A. Ramírez
PY - 2026
DA - 2026/11/03
JO - PLOS One
DO - 10.1371/journal.pone.0347446
UR - https://doi.org/10.1371/journal.pone.0347446
AB - Potato (Solanum tuberosum L.) is a staple crop crucial to global food security, yet its production is severely threatened by late blight (LB), caused by Phytophthora infestans, one of the most destructive plant diseases worldwide. Breeding programs for LB resistance have traditionally relied on labor-intensive and subjective visual assessments, which limit scalability and consistency, particularly in early-generation trials. Unmanned aerial vehicle (UAV)-based remote sensing combined with machine learning (ML) offers a promising alternative for objective, high-throughput disease phenotyping. This study evaluated the potential of this technology to estimate LB severity at the genotype level across two large potato trials in Peru comprising 2,745 breeding clones and 492 genebank accessions. We compared vegetation index–based approaches with a ML framework that integrates K-means clustering and Kernel Ridge Regression (KRR). While NDVI correlated strongly with visually assessed LB severity, particularly at advanced stages, the multispectral KRR approach outperformed linear NDVI models by capturing nonlinear relationships between spectral response and disease progression. Best linear unbiased estimates (BLUEs) of LB severity from both models were strongly and consistently related to the area under the disease progress curve (AUDPC), particularly in later UAV acquisitions. Selection coincidence with AUDPC-based rankings was substantially higher at intermediate to advanced stages of disease progression, indicating that a single UAV inspection at those stages captures enough phenotypic variation to distinguish genotypes. UAV-based multispectral phenotyping integrated with ML, therefore offers a practical, scalable alternative to the time-consuming visual assessments traditionally used in potato breeding programs.
SN - 1932-6203
ER -