Generalized Estimating Equations for Genetic Association Studies of Multi-Correlated Longitudinal Family Data

Abstract

In genetic epidemiology studies, many diseases are multifactorial that can be both environmental and genetic inherited pattern. The relationship between genetic variability and individual phenotypes is usually investigated by genetic association studies. In genetic association studies, longitudinal measures are very important scale in detecting disease variants. They enable to observe both factors in the progress of disease. Generalized Linear Modelling (GLM) techniques offer a flexible approach for testing and quantifying genetic associations considering different types of phenotype distributions. In this study, it is aimed to accommodate Generalized Estimating Equations (GEE) method for genetic association studies in the presence of both familial and serial correlation. For this purpose, a real genotyped data set with the pedigree information and a continuous trait measured over time is used to model the association between the disease and the genotype by analyzing several variants, which have been associated with the disease. A joint working correlation structure is adapted, accounting for two different sources of correlations for estimating equations.