Machine learning-based meta-analysis reveals gut microbiome alterations associated with Parkinson’s disease
Problem
There was no consensus on which gut microbiome features are associated with Parkinson's disease or how well they can diagnose it across multiple studies.
Why it matters
Reliable microbiome-based markers could aid Parkinson's disease diagnosis and treatment.
What the AI did
Machine learning models classified Parkinson's patients from microbiome data across 4489 samples, and meta-analysis of shotgun metagenomes identified PD-associated microbial pathways.
Results
Average AUC 71.9% within studies; 61% across other studies; 68% LOSO with multi-dataset training
Impact
Shows microbiome-based ML classifiers are study-specific and only moderately generalizable for PD diagnosis.
Record
| Venue | Nature Communications |
|---|---|
| Published | 2025 |
| Work type | article |
| Authors | Stefano Romano, Jakob Wirbel, Rebecca Ansorge, Christian Schudoma, Quinten R. Ducarmon, Arjan Narbad, Georg Zeller |
| Citations | 61 (OpenAlex, at import time) |
| References | 107 |
| Model or method | not stated in source |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1038/s41467-025-56829-3 |
| Open access | gold |
| Keywords | machine learning, meta-analysis, gut microbiome, Parkinson's disease, diagnostics |
Importance rating
significant — thousands to millions affected
Parkinson's diagnosis and treatment via microbiome markers; abstract cites strong interest and unprecedented sample scale but states no disease burden numbers; major neurodegenerative disease relevance is inferred.