Entropy-Based Epistasy Search in SNP Case-Control Studies

摘要

The purpose of gene mapping is to identify the causal genetic regions of a specific phenotype mainly a complex disease. Most complex diseases are believed to have multiple contributing loci often having subtle patterns which make them fairly difficult to find in large datasets. We present and discuss a new criterion called conditional mutual information for association mapping and compare it to the previous criterion which is mutual information from different aspects. Furthermore, algorithms are proposed to find relevance chains. The proposed algorithms are especially in favor of diseases having almost equally contributing regions known as being epistatic. These algorithms are applied to both simulated and real data. The real data represents the genotype-phenotype values for AMD disease. Proposed relevance-chain algorithms have detected some highly associated markers with AMD. C# source files for relevance- chains algorithm are freely available at https://www.sharemation.com/amanzour.