Statistics and Its Interface

Volume 7 (2014)

Number 1

We dedicate this special issue to Dr. Gang Zheng, a great colleague and dear friend to many of us.

Genotype-based association models of complex diseases to detect gene-gene and gene-environment interactions

Pages: 51 – 60

DOI: https://dx.doi.org/10.4310/SII.2014.v7.n1.a6

Authors

Iryna Lobach (Department of Neurology, School of Medicine, University of California, San Francisco, Calif., U.S.A.)

Ruzong Fan (Biostatistics and Bioinformatics Branch, Division of Intramural Population Health Research, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Rockville, Maryland, U.S.A.)

Prashiela Manga (Ronald O. Perelman Department of Dermatology, School of Medicine, New York University, New York, N.Y., U.S.A.)

Abstract

A central problem in genetic epidemiology is to identify and rank genetic markers involved in a disease. Complex diseases, such as cancer, hypertension, diabetes, are thought to be caused by an interaction of a panel of genetic factors, that can be identified by markers, which modulate environmental factors. Moreover, the effect of each genetic marker may be small. Hence, the association signal may be missed unless a large sample is considered, or a priori biomedical data are used. Recent advances generated a vast variety of a priori information, including linkage maps and information about gene regulatory dependence assembled into curated pathway databases. We propose a genotype-based approach that takes into account linkage disequilibrium (LD) information between genetic markers that are in moderate LD while modeling gene-gene and gene-environment interactions. A major advantage of our method is that the observed genetic information enters a model directly thus eliminating the need to estimate haplotype-phase. Our approach results in an algorithm that is inexpensive computationally and does not suffer from bias induced by haplotype-phase ambiguity. We investigated our model in a series of simulation experiments and demonstrated that the proposed approach results in estimates that are nearly unbiased and have small variability. We applied our method to the analysis of data from a melanoma case-control study and investigated interaction between a set of pigmentation genes and environmental factors defined by age and gender. Furthermore, an application of our method is demonstrated using a study of Alcohol Dependence.

2010 Mathematics Subject Classification

60K35

Full Text (PDF format)

Published 8 April 2014