Genetic data consists of a wide range of marker types, including common, low-frequency, and rare variants. Multiple genetic markers and their interactions play central roles in the heritability of complex disease. In this study, we propose an algorithm that uses a stratified variable selection design by genetic architectures and interaction effects, achieved by a dataset-adaptive W-test. The polygenic sets in all strata were integrated to form a classification rule. The algorithm was applied to the Critical Assessment of Genome Interpretation 4 bipolar challenge sequencing data. The prediction accuracy was 60% using genetic markers on an independent test set. We found that epistasis among common genetic variants contributed most substantially to prediction precision. However, the sample size was not large enough to draw conclusions for the lack of predictability of low-frequency variants and their epistasis.
This study proposed to perform complex trait prediction using a stratified design. The genetic data are divided into strata according to genetic architectures, and feature selection is conducted within each strata through a data adaptive W-test for main effect and pairwise interactions. An ensemble classification algorithm can be applied to integrate the selected features to perform prediction. Application on CAGI data set showed that including interaction effect of common variants improved prediction accuracy.