Detecting multi-way epistasis in family-based association studies. (13th May 2016)

• Record Type: Journal Article
• Authors: Loucoubar, Cheikh; Grant, Audrey V; Bureau, Jean-François; Casademont, Isabelle; Bar, Ndjido Ardo; Bar-Hen, Avner; Diop, Mamadou; Faye, Joseph; Sarr, Fatoumata Diene; Badiane, Abdoulaye; Tall, Adama; Trape, Jean-François; Cliquet, Freddy; Schwikowski, Benno; Lathrop, Mark; Paul, Richard Edward; Sakuntabhai, Anavaj
• Is Part Of: Briefings in bioinformatics. Volume 18:Number 3(2017:May)
• Page Start: 394
• Page End: 402
• Publication Date: 2016-05-13
• Subjects: family-based genetic association studies -- multi-locus -- epistasis -- interaction -- malaria -- innate immunity; Genetics -- Data processing -- Periodicals; Molecular biology -- Data processing -- Periodicals; Genomes -- Data processing -- Periodicals; 572.80285
• DOI: 10.1093/bib/bbw039
• Languages: English
• ISSNs: 1467-5463

Abstract

Abstract: The era of genome-wide association studies (GWAS) has led to the discovery of numerous genetic variants associated with disease. Better understanding of whether these or other variants interact leading to differential risk compared with individual marker effects will increase our understanding of the genetic architecture of disease, which may be investigated using the family-based study design. We present M-TDT (the multi-locus transmission disequilibrium test), a tool for detecting family-based multi-locus multi-allelic effects for qualitative or quantitative traits, extended from the original transmission disequilibrium test (TDT). Tests to handle the comparison between additive and epistatic models, lack of independence between markers and multiple offspring are described. Performance of M-TDT is compared with a multifactor dimensionality reduction (MDR) approach designed for investigating families in the hypothesis-free genome-wide setting (the multifactor dimensionality reduction pedigree disequilibrium test, MDR-PDT). Other methods derived from the TDT or MDR to investigate genetic interaction in the family-based design are also discussed. The case of three independent biallelic loci is illustrated using simulations for one- to three-locus alternative hypotheses. M-TDT identified joint-locus effects and distinguished effectively between additive and epistatic models. We showed a practical example of M-TDT based on three genes already known to be implicated in malaria susceptibility. Our findings demonstrate the value of M-TDT in a hypothesis-driven context to test for multi-way epistasis underlying common disease etiology, whereas MDR-PDT-based methods are more appropriate in a hypothesis-free genome-wide setting.