Gene-gene Interaction Analysis for Next-generation Sequencing. J. Zhao1, Y. Zhu1, M. Xiong2 1) Department of Epidemiology, Tulane University School of Public Health and Tropical Medicine, New Orleans, LA; 2) Human Genetics Center, the University of Texas School of Public Health, Houston, TX.
Traditional gene-gene interaction analysis was originally designed to test pair-wise genetic interactions among common variants. These methodologies are difficult to apply to rare variants because of prohibitive computational time, large number of tests and low statistical power. Rare variants generated by next-generation sequencing (NGS) pose great challenges for genetic interaction analysis due to the following reasons: (1) the demands in the paradigm of changes in interaction analysis; (2) the severe multiple-testing problems, and (3) the expensive computations. To meet these challenges, here we propose a novel statistical method that shifts the paradigm from interaction between two SNPs to the interaction between two genomic regions. In other words, we treat a gene, instead of a SNP, as the unit of analysis and use functional data analysis techniques as dimensional reduction tools to collectively test interactions between all possible SNP pairs within two genomic regions, including both common and rare variants. Through intensive simulation analyses, we demonstrated that this novel approach has correct type 1 error rates and higher power in detecting genetic interactions compared to other existing methodologies. The proposed statistic was applied to several real NGS datasets of cardiovascular disease, including the Wellcome Trust Case Control Consortium (WTCCC) study, the Framingham Heart Study (FHS), and the NHLBIs Exome Sequencing Project. Of the 27 significantly interacting gene pairs identified in the FHS, 6 interacting pairs were able to be replicated in the WTCCC study and 24 pairs were able to be confirmed in the EOMI study after accounting for multiple testing by Bonferroni correction, indicating that the proposed novel statistic has a great potential in genetic interaction analysis for NGS data.
You may contact the first author (during and after the meeting) at