Related Experiment Video
Updated: May 17, 2025
Author Spotlight: Exploring Non-Motor Symptoms in Parkinson's Disease
Published on: September 22, 2023
Machine learning-based meta-analysis reveals gut microbiome alterations associated with Parkinson's disease
Stefano Romano1,2, Jakob Wirbel3, Rebecca Ansorge4,5
1Quadram Institute Bioscience, Norwich Research Park, Norwich, UK. stfno.rmno@gmail.com.
Abstract:
There is strong interest in using the gut microbiome for Parkinson's disease (PD) diagnosis and treatment. However, a consensus on PD-associated microbiome features and a multi-study assessment of their diagnostic value is lacking. Here, we present a machine learning meta-analysis of PD microbiome studies of unprecedented scale (4489 samples). Within most studies, microbiome-based machine learning models accurately classify PD patients (average AUC 71.9%). However, these models are study-specific and do not generalise well across other studies (average AUC 61%). Training models on multiple datasets improves their generalizability (average LOSO AUC 68%) and disease specificity as assessed against microbiomes from other neurodegenerative diseases. Moreover, meta-analysis of shotgun metagenomes delineates PD-associated microbial pathways potentially contributing to gut health deterioration and favouring the translocation of pathogenic molecules along the gut-brain axis. Strikingly, microbial pathways for solvent and pesticide biotransformation are enriched in PD. These results align with epidemiological evidence that exposure to these molecules increases PD risk and raise the question of whether gut microbes modulate their toxicity. Here, we offer the most comprehensive overview to date about the PD gut microbiome and provide future reference for its diagnostic and functional potential.
Insights
Machine learning models show promise for diagnosing Parkinson's disease (PD) using gut microbiome data. However, models trained on single studies lack generalizability, highlighting the need for multi-study approaches for reliable PD microbiome analysis.
Area of Science:
- Microbiome research
- Neurodegenerative diseases
- Machine learning applications
Background:
- Growing interest in the gut microbiome for Parkinson's disease (PD) diagnosis and treatment.
- Lack of consensus on PD-associated microbiome features and their diagnostic value across studies.
- Need for large-scale, multi-study assessment of the PD gut microbiome.
Purpose of the Study:
- To conduct a large-scale machine learning meta-analysis of PD microbiome studies.
- To assess the generalizability and diagnostic accuracy of PD microbiome signatures.
- To identify PD-associated microbial pathways and their role in disease pathogenesis.
Main Methods:
- Machine learning meta-analysis of 4489 gut microbiome samples from PD patients and controls.
- Development and validation of microbiome-based classification models.
- Shotgun metagenomic analysis to identify microbial pathways associated with PD.
Main Results:
- Microbiome models achieved good accuracy within individual studies (AUC 71.9%) but poor generalizability across studies (AUC 61%).
- Multi-dataset training improved model generalizability (LOSO AUC 68%) and disease specificity.
- Enrichment of microbial pathways for solvent and pesticide biotransformation observed in PD, potentially linking environmental exposures to disease risk.
Conclusions:
- The gut microbiome holds diagnostic potential for Parkinson's disease, but robust models require multi-study data.
- Gut microbial pathways may influence PD pathogenesis and interact with environmental risk factors.
- This comprehensive analysis provides a reference for the diagnostic and functional roles of the PD gut microbiome.
Related Concept Videos
01:15Parkinson's Disease: Overview
01:37Neural Regulation
01:24Parkinson's Disease: Treatment
Parkinson's Disease is primarily a result of the loss of dopaminergic neurons in the substantia nigra pars compacta. The cornerstone of...