Related Experiment Videos
Protocol for Epistasis Detection with Machine Learning Using GenEpi Package
Olutomilayo Olayemi Petinrin1,2, Ka-Chun Wong3
1Department of Computer Science, College of Engineering, City University of Hong Kong, Kowloon Tong, Hong Kong.
Methods in Molecular Biology (Clifton, N.J.)
|March 18, 2021
Summary
Identifying genetic variants associated with disease is crucial for developing new treatments. The GenEpi Python package offers a novel machine learning approach to detect gene interactions (epistasis) and predict disease risk, improving upon existing methods.
Area of Science:
- Genetics
- Bioinformatics
- Computational Biology
Background:
- Identifying genetic variants associated with disease is essential for medical treatment and prevention strategies.
- Genome-wide association studies (GWAS) aim to uncover genetic underpinnings of disease susceptibility.
- Detecting epistasis (gene-gene interactions) is challenging due to limited statistical power and lack of standardized methods.
Purpose of the Study:
- To introduce the GenEpi Python package for detecting within-gene and cross-gene epistasis.
- To provide a standardized machine learning workflow for epistasis detection.
- To facilitate the development of predictive models incorporating genetic and environmental factors.
Main Methods:
- Utilizes a two-level workflow machine learning model.
- Employs a three-step procedure: dimensionality reduction, within-gene epistasis selection, and cross-gene epistasis selection.
- Demonstrates the package usage with example data.
Main Results:
- The GenEpi package effectively detects both within-gene and cross-gene epistasis.
- The package provides a robust framework for analyzing complex genetic interactions.
- Enables the construction of prediction models integrating genetic and environmental influences.
Conclusions:
- GenEpi offers a standardized and powerful tool for epistasis detection in genetic studies.
- The package aids in understanding disease vulnerability by identifying gene-gene interactions.
- Facilitates the development of personalized medicine through integrated genetic and environmental risk prediction.