A non-parametric approach for detecting gene-gene interactions associated with age-at-onset outcomes.

ABSTRACT:

Background

Cox-regression-based methods have been commonly used for the analyses of survival outcomes, such as age-at-disease-onset. These methods generally assume the hazard functions are proportional among various risk groups. However, such an assumption may not be valid in genetic association studies, especially when complex interactions are involved. In addition, genetic association studies commonly adopt case-control designs. Direct use of Cox regression to case-control data may yield biased estimators and incorrect statistical inference.

Results

We propose a non-parametric approach, the weighted Nelson-Aalen (WNA) approach, for detecting genetic variants that are associated with age-dependent outcomes. The proposed approach can be directly applied to prospective cohort

SUBMITTER: Li M 

PROVIDER: S-EPMC4087128 | biostudies-literature | 2014 Jul

REPOSITORIES: biostudies-literature