Deciphering Plant–Microbe Networks through Game Theory and Machine Learning

April 21, 2026

A collaborative research group from The University of Tokyo, RIKEN CSRS, and French National Museum of Natural History has developed a new analytical framework that integrates genomic, metabolomic, and microbiome data to uncover the complex relationships associated with plant traits.
Plant growth and characteristics are influenced not only by genetic information encoded in DNA, but also by microorganisms in the root environment and metabolites produced within the plant. However, because these datasets differ greatly in nature and interact in highly complex ways, achieving an integrated understanding has been challenging.

In this study, the researchers applied machine learning to capture complex nonlinear relationships that conventional linear models could not adequately describe, enabling the visualization of connections among multiple layers of biological information through a novel variable-selection approach. Furthermore, by employing a game theory–based method known as SHapley Additive exPlanations (SHAP), they quantitatively evaluated the contribution of each factor to the model’s predictions. This approach not only improved predictive performance, but also enabled biological interpretation of the underlying mechanisms.

Using soybean datasets, the team compared drought and normal growth conditions and demonstrated that the key contributing factors and their relationships change depending on environmental conditions.

These findings are expected to advance understanding of plant–microbe interactions and contribute to applications in crop improvement and sustainable agriculture.

 

Original article

BMC Environmental Microbiome　doi: 10.1186/s40793-026-00883-x

H. Yoshioka, P. Debeljak, S. Prado, Y. Fuji, Y. Ichihashi, H. Iwata,

"Interpretable multi-omics machine learning reveals drought-driven shifts in plant–microbe interactions".

Contact

Yushiro Fuji; Research Scientist
Metabolic Systems Research Team
Yasunori Ichihashi; Team Director
Holobiont and Resilience Research Team