Article: Multifactor dimensionality reduction analysis of multiple binary traits for gene-gene interaction Journal: International Journal of Data Mining and Bioinformatics (IJDMB) 2016 Vol.14 No.4 pp.293 - 304 Abstract: Beyond single variant analysis of genome-wide association studies (GWAS), joint identification approaches including gene-gene interaction analysis have been studied vigorously and many novel results have been obtained. However, most GWAS have been focused only on a single trait. Since many complex diseases that result from damage to biological pathways can be pleiotropic, a traditional univariate analysis focusing on a single trait may not detect the full genetic architecture of complex diseases well. Therefore, in order to improve power and reflect comprehensive biological associations, we investigated a multivariate approach for analysing multiple traits simultaneously. For gene-gene interaction analysis, we extended a multifactor dimensionality reduction (MDR) approach to handle multiple traits. From simulation studies, we confirmed that the multivariate approach provides more stable, and precise accuracy measures compared to univariate analysis. We applied our approach to a GWA longitudinal dataset of 8842 Korean individuals and detected genetic variants associated with hypertension and hyperlipidemia using multiple traits. Inderscience Publishers - linking academia, business and industry through research

Title: Multifactor dimensionality reduction analysis of multiple binary traits for gene-gene interaction

Authors: Iksoo Huh; Taesung Park

Addresses: Department of Statistics, Seoul National University, Gwanak-gu, Seoul 151-747, South Korea ' Department of Statistics, Seoul National University, Gwanak-gu, Seoul 151-747, South Korea

Abstract: Beyond single variant analysis of genome-wide association studies (GWAS), joint identification approaches including gene-gene interaction analysis have been studied vigorously and many novel results have been obtained. However, most GWAS have been focused only on a single trait. Since many complex diseases that result from damage to biological pathways can be pleiotropic, a traditional univariate analysis focusing on a single trait may not detect the full genetic architecture of complex diseases well. Therefore, in order to improve power and reflect comprehensive biological associations, we investigated a multivariate approach for analysing multiple traits simultaneously. For gene-gene interaction analysis, we extended a multifactor dimensionality reduction (MDR) approach to handle multiple traits. From simulation studies, we confirmed that the multivariate approach provides more stable, and precise accuracy measures compared to univariate analysis. We applied our approach to a GWA longitudinal dataset of 8842 Korean individuals and detected genetic variants associated with hypertension and hyperlipidemia using multiple traits.

Keywords: multifactor dimensionality reduction; GWAS; genome-wide association studies; gene-gene interaction; bioinformatics; multiple binary traits; biological associations; multiple traits; simulation; South Korea; genetic variants; hypertension; hyperlipidemia.

DOI: 10.1504/IJDMB.2016.075810

International Journal of Data Mining and Bioinformatics, 2016 Vol.14 No.4, pp.293 - 304

Received: 31 Jan 2015
Accepted: 31 Jan 2015
Published online: 06 Apr 2016 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: Multifactor dimensionality reduction analysis of multiple binary traits for gene-gene interaction

Keep up-to-date