GEP-EpiSeeker: a gene expression programming-based method for epistatic interaction detection in genome-wide...

Area of Science:

• Genetics and Bioinformatics
• Computational Biology
• Statistical Genetics

Background:

• Epistatic interactions between single nucleotide polymorphisms (SNPs) are crucial for understanding complex diseases.
• Identifying these interactions is challenging due to the combinatorial explosion of SNP combinations.

Purpose of the Study:

• To develop a novel, efficient, and accurate method for detecting epistatic interactions.
• To address the computational challenges in analyzing large-scale SNP data for disease association studies.

Main Methods:

• Formulated epistasis detection as a combinatorial optimization problem.
• Developed GEP-EpiSeeker, an evolutionary framework utilizing Gene Expression Programming.
• Incorporated Bayesian network-based fitness evaluation and Chi-square tests for SNP combination identification.
• Designed specialized genetic operators and adaptive manipulation methods to enhance evolutionary efficiency and accuracy.

Main Results:

• GEP-EpiSeeker demonstrated superior performance compared to state-of-the-art methods (BEAM, BOOST, AntEpiSeeker, MACOED, EACO).
• Evaluated on GWAS datasets for 12 DME and 10 DNME disease models, showing improvements in power, recall, precision, and F1-score.
• Successfully identified significant epistatic interactions within complex disease models.

Conclusions:

• GEP-EpiSeeker offers a novel and effective approach for epistasis detection in Genome-wide Association Studies.
• The method provides a promising alternative for accurately identifying complex genetic interactions underlying diseases.
• This framework advances the systematic exploration of SNP associations with complex traits.