A Novel Statistic for Genome-Wide Interaction Analysis

• DOI: 10.1371/journal.pgen.1001131
• Publication Date: 2010-09-23T19:38:27Z
• Authors: Xuesen Wu; Hua Dong; Li Luo; Yun Zhu; Gang Peng; John D. Reveille; Momiao Xiong
• Keywords: Missing heritability problem; Genome-wide Association Study; Statistic; Imputation (statistics); Genetic Association; Statistical power

Although great progress in genome-wide association studies (GWAS) has been made, the significant SNP associations identified by GWAS account for only a few percent of genetic variance, leading many to question where and how we can find missing heritability. There is increasing interest interaction analysis as possible source finding heritability unexplained current GWAS. However, existing statistics testing have low power analysis. To meet challenges raised interactional analysis, developed novel statistic between two loci (either linked or unlinked). The null distribution type I error rates new are validated using simulations. Extensive show that much higher detect than classical logistic regression. results 44 211 pairs SNPs showing evidence interactions with FDR<0.001 0.001<FDR<0.003, respectively, which were seen independent psoriasis. These included five interacting genes LST1/NCR3, CXCR5/BCL9L, GLS2, some located target sites miR-324-3p, miR-433, miR-382, well 15 had nonsynonymous substitutions. Our demonstrated valuable tool remaining GWAS, able search across genome. Real data showed be replicated studies.