Black cottonwood is a fast-growing, genetically diverse tree with potential as a lignocellulosic feedstock for biofuels, including sustainable aviation fuels. Yet breeding improved trees depends on connecting genetic variation with complex traits such as biomass, stress tolerance, and photosynthetic performance. Conventional tree phenotyping is slow, labor-intensive, often destructive, and poorly suited to tracking tall plants repeatedly through changing field conditions. Hyperspectral remote sensing can capture biochemical and structural information rapidly, while genome-wide association studies can locate genetic variants linked to those traits, but their integration has rarely been tested continuously in field-grown trees. Because of these challenges, deeper investigation of time-resolved genotype–phenotype relationships is needed.

Researchers from the Department of Plant Sciences at the University of California, Davis, and the Biosciences Division and Center for Bioenergy Innovation at Oak Ridge National Laboratory reported (DOI: 10.1016/j.plaphe.2026.100174) the study online on March 27, 2026, in Plant Phenomics. The team combined tower-based hyperspectral sensing with genome-wide association studies (GWAS) to examine 1,423 wavebands from 400 to 900 nanometers across 505 Populus trichocarpa genotypes, asking how spectral traits are inherited, how their genetic architecture changes over time, and how well they predict growth.

The researchers monitored 595 trees representing 505 genotypes in a drought-treated field plot, using a 15-meter tower system to repeatedly sample canopy reflectance and solar-induced chlorophyll fluorescence. They calculated the normalized difference vegetation index (NDVI), photochemical reflectance index (PRI), chlorophyll:carotenoid index (CCI), and solar-induced fluorescence (SIF), then treated daily spectral features as phenotypes for GWAS. Heritability was strongest in blue and red regions linked to chlorophyll and carotenoid absorption; CCI and NDVI showed broad-sense heritability (H²) of roughly 0.20–0.25. High-confidence filtering identified 98 single nucleotide polymorphism (SNP) associations within the 400–700-nanometer range, with substantial overlap between blue- and red-associated signals. More than 10 candidate genes were linked to chloroplast function, including Potri.017G109600, Potri.005G075700, Potri.012G080800, Potri.001G361800, and Potri.005G012600. Daily GWAS across four vegetation indices produced 272 analyses over 68 days and showed that nearby measurement dates had more similar genetic effects, indicating that trait architecture shifted as trees moved from active growth toward senescence. Spectral traits also tracked relative growth rate, while an elastic-net model using multiple wavelengths predicted genotype growth rankings more accurately than any single band or vegetation index, reaching Spearman’s ρ above 0.5.

The authors said the study’s central message is that plant genetics can leave measurable signatures in canopy reflectance, but those signatures are not fixed through the season. By following the same trees repeatedly, the team could see both where genetic control was strongest and when particular associations became more or less informative. They said this makes timing an important part of high-throughput phenotyping: a measurement that is useful during peak growth may tell a different genetic story during drought progression or autumn senescence, especially for pigment- and photosynthesis-related traits.

The findings point toward faster screening strategies for tree improvement, especially when breeders must evaluate large populations under realistic field stress. Time-series hyperspectral measurements could help identify promising genotypes earlier, reduce reliance on destructive measurements, and guide selection for productivity or stress resilience. The authors also propose testing the framework across multiple years, environments, and larger genetic panels to determine how robust the detected signals are. Extending measurements to calibrated airborne or unmanned aerial vehicle (UAV) platforms could increase scale, while feature selection may reveal a smaller set of informative wavelengths for lower-cost multispectral sensors. Experimental validation will still be needed before candidate genes or spectral markers are used routinely in breeding.

###

References

DOI

10.1016/j.plaphe.2026.100174

Original Source URL

https://doi.org/10.1016/j.plaphe.2026.100174

Funding information

This work was supported by the U.S. Department of Energy under Contract No. DE-AC05-00OR22725 and the Genomics enabled Plant Biology for Determination of Gene Function programme (DE-SC0020164). Additional support was provided by the James Monroe McDonald Endowment, the John B. Orr Chair Endowment, and University College London.

About Plant Phenomics

Plant Phenomics publishes breakthrough research that integrates genomics, genetics, physiology, molecular biology, bioinformatics, statistics, mathematics and computer science to advance plant phenotyping—from the cellular level to entire populations. The journal tackles key scientific challenges in phenomics by showcasing cuttingedge technologies for data acquisition, management, interpretation and modelling. Its scope spans practical applications across all plant sciences, aiming to translate novel phenotyping tools into realworld solutions for breeding, crop management and fundamental plant biology. By fostering interdisciplinary innovation, Plant Phenomics serves as a hub for researchers developing highthroughput, accurate and scalable methods to decipher complex plant traits, ultimately accelerating progress in agriculture and plant science.