Related Experiment Video
Updated: Dec 17, 2025
08:04
Identification and Classification of Position-specific GABAA Receptor Subunit Missense Variants for Their Role In Hippocampal Pyramidal Neurons
Published on: June 6, 2025
1.2K
A Knowledge-Based Machine Learning Approach to Gene Prioritisation in Amyotrophic Lateral Sclerosis
Daniel M Bean1,2, Ammar Al-Chalabi3,4, Richard J B Dobson1,2,5
1Department of Biostatistics & Health Informatics, King's College London, 16 De Crespigny Park, London SE5 8AF, UK.
Genes
|June 25, 2020
Summary
This study introduces a machine learning approach to identify novel genes linked to amyotrophic lateral sclerosis (ALS). The method successfully predicted genes associated with neurodegenerative diseases, advancing our understanding of ALS genetics.
Area of Science:
- Genetics
- Neuroscience
- Computational Biology
Background:
- Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disease primarily affecting motor neurons, with genetic factors playing a role in approximately 15% of cases.
- Current knowledge of ALS genetic landscape is limited, necessitating advanced methods for novel gene discovery.
- Protein-protein interaction and phenotype-genotype data offer a foundation for machine learning models to predict new candidate genes.
Purpose of the Study:
- To develop and validate a knowledge-based machine learning method for predicting novel candidate genes associated with amyotrophic lateral sclerosis (ALS).
- To investigate the biological relevance and potential diagnostic value of the predicted ALS genes.
Main Methods:
- Trained a machine learning model using protein-protein interaction data (IntAct), gene function annotations (Gene Ontology), and known disease-gene associations (DisGeNet).
- Input known ALS genes from public databases and manual review to generate lists of novel candidate genes.
- Evaluated candidate gene relevance using genome-wide association study (GWAS) summary statistics and performed functional/phenotype enrichment analysis.
Main Results:
- Predicted gene sets showed enrichment for genes associated with other neurodegenerative diseases genetically and phenotypically overlapping with ALS.
- Enrichment analysis revealed associations with biological processes relevant to ALS.
- Using ALS genes from ClinVar and manual review as input, predicted sets were significantly enriched for ALS-associated genes in GWAS prioritization (ClinVar p = 0.038, manual review p = 0.060).
Conclusions:
- The developed machine learning method effectively predicts novel candidate genes for amyotrophic lateral sclerosis (ALS).
- Predicted genes are biologically relevant and show genetic and phenotypic overlap with other neurodegenerative diseases.
- This approach enhances gene prioritization in ALS genome-wide association studies, contributing to a deeper understanding of ALS genetics.
Keywords:
amyotrophic lateral sclerosisgene discoverygene prioritisationknowledge graphmachine learningmotor neurone disease
