A genome-wide screen of gene–gene interactions for rheumatoid arthritis susceptibility

Abstract

The objective of the study was to identify interacting genes contributing to rheumatoid arthritis (RA) susceptibility and identify SNPs that discriminate between RA patients who were anti-cyclic citrullinated protein positive and healthy controls. We analyzed two independent cohorts from the North American Rheumatoid Arthritis Consortium. A cohort of 908 RA cases and 1,260 controls was used to discover pairwise interactions among SNPs and to identify a set of single nucleotide polymorphisms (SNPs) that predict RA status, and a second cohort of 952 cases and 1,760 controls was used to validate the findings. After adjusting for HLA-shared epitope alleles, we identified and replicated seven SNP pairs within the HLA class II locus with significant interaction effects. We failed to replicate significant pairwise interactions among non-HLA SNPs. The machine learning approach "random forest" applied to a set of SNPs selected from single-SNP and pairwise interaction tests identified 93 SNPs that distinguish RA cases from controls with 70% accuracy. HLA SNPs provide the most classification information, and inclusion of non-HLA SNPs improved classification. While specific gene-gene interactions are difficult to validate using genome-wide SNP data, a stepwise approach combining association and classification methods identifies candidate interacting SNPs that distinguish RA cases from healthy controls.

Overview of random forest methodology and practical guidance with emphasis on computational biology and bioinformatics

“…In the application of RFs to genome-wide association data, the focus has been on different features of the algorithm. Whereas some used RFs to identify candidate regions similar to standard analyses [31], others focused on the detection of gene-gene interactions [46]. In a third group of applications, the resulting genetic regions are not of interest in themselves; instead, a prediction model is built using hundreds of SNPs at a time [19].…”

A Review for Detecting Gene-Gene Interactions Using Machine Learning Methods in Genetic Epidemiology

“…The method of RF used by Liu et al [36] is to identify rheumatoid arthritis (RA) susceptibility by contributed gene-gene interactions and to identify SNPs which can distinguish who were anticyclic citrullinated protein positive and healthy controls among RA patients. The result showed that RF had distinguish RA cases from controls with 70% accuracy when applied to a set of SNPs selected from single-SNP and pairwise interaction tests identified 93 SNPs.…”

Mining massive SNP data for identifying associated SNPs and uncovering gene relationships

“…For example, SVM, random forest, and neural network methods have been used to identify SNP combinations that are associated with type 2 diabetes [2], childhood allergic asthma [28], rheumatoid arthritis [17], and other diseases [24]. However, these methods need to focus on a small set of candidate SNPs, often on candidate disease genes, in order to identify SNP combinations for targeted diseases or phenotypes.…”