Multi-omic integration using interpretable machine learning reveals genetic mechanisms of trait variation and phenotypic plasticity in switchgrass

Explainable AI models identified key genes and interactions predicting flowering time and biomass in switchgrass across diverse environments, revealing both known and novel regulators.

• Why it matters: Understanding the genetic basis of complex traits like flowering time and biomass is difficult due to small gene effects and environmental interactions, limiting crop improvement efforts.
• What they did: Researchers integrated genome-wide SNPs and RNA-seq data from a switchgrass panel grown in Michigan and Texas, applying interpretable machine learning models to predict phenotypic variation and plasticity.
• The result: The approach uncovered important genes, including canonical regulators and environment-specific candidates, and revealed gene-gene interactions, enabling mechanistic hypotheses for trait control and guiding breeding strategies.