Genome-wide association studies for bivariate sparse longitudinal data.

ABSTRACT:

Objective

Longitudinal measurements with bivariate response have been analyzed by several authors using two separate models for each response. However, for most of the biological or medical experiments, the two responses are highly correlated and hence a separate model for each response might not be a desirable way to analyze such data. A single model considering a bivariate response provides a more powerful inference as the correlation between the responses is modeled appropriately. In this article, we propose a dynamic statistical model to detect the genes controlling human blood pressure (systolic and diastolic).

Methods

By modeling the mean function with orthogonal Legendre polynomials and the covariance matrix with a stationary parametric structure, we incorporate the s

SUBMITTER: Das K 

PROVIDER: S-EPMC3725885 | biostudies-literature | 2011

REPOSITORIES: biostudies-literature