Foliar infrared spectra track nematode density and symptom-specific phytobiome signatures in beech leaf disease

Beech leaf disease (BLD), caused by the foliar nematode Litylenchus
crenatae ssp. mccannii (LCM) is an emerging threat to American beech
(Fagus grandifolia) across eastern North America. In this study,
near-infrared (NIR) spectroscopy was integrated with microbiome sequencing
and machine learning approaches to evaluate whether host spectral
signatures can predict pathogen abundance and associated phytobiome
composition. Spectral profiles were analyzed using multivariate
statistical approaches and predictive models, including random forest,
support vector machines, elastic net regression, and partial least squares
regression. NIR spectra accurately predicted nematode abundance and
reconstructed major gradients of bacterial community differentiation
derived from 16S rRNA gene sequencing, with nonlinear models exhibiting
the highest predictive performance. These results indicate that host
spectral signatures encode biochemical information reflecting both
pathogen burden and microbiome structure. Our findings demonstrate the
potential of NIR spectroscopy as a rapid, non-destructive tool for
characterizing host–pathogen–microbiome interactions and improving
detection of BLD in forest ecosystems.