1121- Jean Shin

Inferring gene-environment interaction from case-parent trio data: Evaluation of and adjustment for spurious GxE and development of a data-smoothing method to uncover true GxE

 

Most complex diseases are influenced jointly by genes (G) and environmental or non-genetic attributes (E). Gene-environment interaction (GxE) is measured by statistical interaction between G and E, which occurs when genotype relative risks (GRRs) vary with E. In this thesis, we explore the sources of spurious GxE and propose a data-smoothing approach to GxE for case-parent trio data.

In the first project, we address the problem of making inference about GxE based on transmission rates. Since GRRs that vary with E lead to transmission rates that do too, transmission rates have been used to make inference about GxE. However transmission-based tests of GxE are found to be invalid in general under population stratification. To better understand the bias, we derive theoretical transmission rates and, comparing them to GRRs in a range of settings and through simulation, we investigate the practical implication of the bias.

Although a number of parametric and non-parametric approaches to GxE in case-parent trio data are robust against population stratification, they assume or require specifying a parametric model for GxE. Consequently, model mis-specification can lead to a loss of statistical power. In the second project, we develop a data-smoothing method to explore GxE that does not require model specification for the interaction component. The data-driven method produces graphical displays of GxE that suggest their forms, using a generalized additive modelling framework. For testing significance of GxE, we take a permutation approach to account for the additional uncertainty introduced by the smoothing process.

For many approaches to inference of GxE with case-parent trio data, including our own, a key assumption is that the test marker is causal; however, in reality, it may not be causal but linked to a causal gene. In this case, the approaches can give a false impression of GxE due to a form of population stratification that has not been appreciated well. In the final project, we investigate, through simulation, the source of the spurious GxE and propose an adjustment that uses additional unlinked markers on the affected offspring.