Research On Genome-wide Single Nucleotide Polymorphisms' Interaction Detection Methods

Genome-wide association studies(GWAS)have been widely used in complex disease study.However,the identified single risk SNPs in GWAS can only explain a portion of the theoretical estimated heritability of complex diseases.It is widely acknowledged that interaction effects of multiple SNPs may also uncover a portion of unexplained heritability of complex diseases.How to detect genome-wide SNP interactions has been attracting more and more attention.In population-based and family-based SNP interaction studies,many efficient methods have been proposed.However,the existing methods still suffer from some issues,such as intensive computational burden,ignore high-order SNP interactions and do not allow marginal effect adjustment.To combat with these issues,this dissertation conducted an in-depth study on genome-wide interaction detection and the specific works of this paper are outlined as below:(1)Research on case-control based high-order SNP interaction detection.Many case-control based interaction detection methods have been proposed,but most of these methods only focus on pairwise SNP interactions and ignore high-order SNP interactions.High-order interactions also contribute to variability in complex diseases.Existing methods for high-order interaction detection can hardly handle genome-wide data and suffer from low detection power.In this paper,we proposed a flexible two-stage approach(called HiSeeker)to detect high-order interactions.In the screening stage,in order to reduce the search space and retain more valid information,HiSeeker employs the chi-squared test and logistic regression model to efficiently obtain candidate pairwise combinations,which have intermediate or significant associations with the disease for interaction detection.In the search stage,according to the number of candidate combinations,HiSeeker chooses different search strategies(exhaustive search or ant colony optimization-based search)to detect high-order interactions from candidate combinations.The flexible search strategy allows HiSeeker to identify as much interaction as possible while ensuring computational efficiency.The experimental results on simulated datasets demonstrate that HiSeeker can more efficiently and effectively detecting high-order interactions than related representative algorithms.On two real GWAS datasets,HiSeeker detects several significant high-order interactions,and these high-order interactions can hardly be identified by related algorithms.This result also demonstrates that HiSeeker is utility and effective.(2)Research on trio-based SNP interaction detection.Compared with population-based interaction detection methods,family-based methods are robust to population stratification.Due to the difficulty of collecting family data and the complexity of different types of pedigree structure,few researchers focus on family-based interaction detection.The existing family-based methods suffer from low detection power and heavy computational burden.Furthermore,they do not allow for marginal effect adjustment.In this paper,we combine Multifactor Dimensionality Reduction(MDR)and regressing methods,propose a method(TrioMDR)to detect SNP interaction in trio families.TrioMDR combines the MDR with logistic regression models to check interactions,which allow TrioMDR can adjust marginal effects.In addition,unlike consuming permutation procedures used in traditional MDR-based methods,TrioMDR utilizes a simple semi-parameter P-values correction procedure to control type I error rate,this procedure only uses several permutations to achieve the significance of a multi-locus model and significantly speedups TrioMDR.We performed extensive experiments on simulated data to evaluate the type I error and power of TrioMDR under different scenarios.The results demonstrate that TrioMDR is fast and more powerful in general than some recently proposed methods for interaction detection in trios.