Research progress in machine learning methods for gene-gene interaction detection.

Area of Science:

• Genetics and Bioinformatics
• Computational Biology
• Statistical Genetics

Background:

• Complex diseases arise from intricate gene-gene and gene-environment interactions.
• Identifying these high-dimensional gene-gene interactions presents significant computational challenges.
• Machine learning (ML) has emerged as a powerful tool for detecting gene-gene interactions over the past two decades.

Purpose of the Study:

• To review and summarize advancements in machine learning methods for gene-gene interaction detection.
• To systematically examine the principles and limitations of current ML techniques applied to genome-wide association studies (GWAS).
• To provide insights into future research directions in this field.

Main Methods:

• Review of machine learning approaches including Neural Networks (NN), Random Forest (RF), Support Vector Machines (SVM), and Multifactor Dimensionality Reduction (MDR).
• Analysis of the application of these methods in the context of genome-wide association studies (GWAS).
• Systematic examination of the underlying principles and inherent limitations of each ML method.

Main Results:

• Machine learning methods have shown success in detecting gene-gene interactions.
• NN, RF, SVM, and MDR are key ML techniques employed in GWAS for interaction detection.
• Current methods have specific strengths and weaknesses that influence their applicability.

Conclusions:

• Machine learning offers promising solutions for the computationally intensive task of detecting gene-gene interactions.
• Further research is needed to refine existing ML methods and explore novel approaches for complex disease genetics.
• Understanding the limitations of current ML techniques is crucial for advancing gene-gene interaction detection in GWAS.