Interpreting graph neural networks reveals microbial association patterns in inflammatory bowel diseases
- 1Technical University of Denmark
Abstract
Traditional machine learning models trained on gut microbiome abundance profiles can distinguish individuals with inflammatory bowel disease (IBD) from non-IBD individuals. However, their interpretation typically considers taxa in isolation, making it difficult to understand how associations between taxa contribute to predictions. Here, we inferred individual-specific microbial association networks from longitudinal metagenomics data of 106 individuals in HMP2/iHMP and used them to train and interpret graph neural network (GNN) models, which we compared with abundance-based models such as Random Forest (RF). Although RF achieved higher predictive performance (AUPRC ~0.59 vs 0.446), interpreting the GNNs revealed that their predictions relied on a densely connected module of co-abundant Bacteroidales species. Unlike the taxa prioritized by RF, these species were not differentially abundant between conditions. In an independent cohort, this module carried most of the predictive signal, and removing its associations reduced performance to near chance. These results show that coordinated association structure, rather than abundance shifts alone, is a feature of the IBD gut microbiome, and that interpreting GNNs can uncover such patterns that complement abundance-based analyses.
Related
- Neuropeptide T2A-GAL4 knock-ins illustrate practical considerations for interpreting GAL4 reporter patterns in Drosophila
- Interpreting Graph Neural Networks for Antibacterial Drug Repurposing against Acinetobacter baumannii
- Interpreting chemical crosslinks: Score-based approaches and deep neural networks
- Interpreting convolutional neural networks in population genetics.
- Exploring the biogeographical patterns and driving mechanisms of microbial communities in four microhabitats of Zostera japonica beds
BibTeX
@misc{cunha2026interpreting,
title = {{Interpreting graph neural networks reveals microbial association patterns in inflammatory bowel diseases}},
author = {A. F. Cunha and E. Vidal and M. Barranco-Altirriba and T. Amigo and J. G. Assis and H. I. Nakaya and A. Santos},
howpublished = {bioRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.27.762128},
url = {https://www.biorxiv.org/content/10.64898/2026.10.27.762128v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Interpreting graph neural networks reveals microbial association patterns in inflammatory bowel diseases
AU - A. F. Cunha
AU - E. Vidal
AU - M. Barranco-Altirriba
AU - T. Amigo
AU - J. G. Assis
AU - H. I. Nakaya
AU - A. Santos
PY - 2026
DA - 2026/10/27
PB - bioRxiv
DO - 10.64898/2026.10.27.762128
UR - https://www.biorxiv.org/content/10.64898/2026.10.27.762128v1
AB - Traditional machine learning models trained on gut microbiome abundance profiles can distinguish individuals with inflammatory bowel disease (IBD) from non-IBD individuals. However, their interpretation typically considers taxa in isolation, making it difficult to understand how associations between taxa contribute to predictions. Here, we inferred individual-specific microbial association networks from longitudinal metagenomics data of 106 individuals in HMP2/iHMP and used them to train and interpret graph neural network (GNN) models, which we compared with abundance-based models such as Random Forest (RF). Although RF achieved higher predictive performance (AUPRC ~0.59 vs 0.446), interpreting the GNNs revealed that their predictions relied on a densely connected module of co-abundant Bacteroidales species. Unlike the taxa prioritized by RF, these species were not differentially abundant between conditions. In an independent cohort, this module carried most of the predictive signal, and removing its associations reduced performance to near chance. These results show that coordinated association structure, rather than abundance shifts alone, is a feature of the IBD gut microbiome, and that interpreting GNNs can uncover such patterns that complement abundance-based analyses.
ER -