Adapting to Unknown Genetic Effects for Novel Gene Detection Using GWAS Summary Data

• Activity Number: 81 - Contributed Poster Presentations: Section on Statistics in Epidemiology
• Type: Contributed
• Date/Time: Monday, August 3, 2020 : 10:00 AM to 2:00 PM
• Sponsor: Section on Statistics in Genomics and Genetics
• Author(s): Xiaohui Chen* and Zheyang Wu
• Companies: and Worcester polytechnic Institute
• Keywords: SNP set test; meta-analysis; p-value combination; genome-wide association studies

Abstract

One important strategy for current Genome-Wide Association Study (GWAS) is to combine information based on many meta-studies that have generated rich data of summary statistics for large number of SNP. Many methods exist from literature, however, most of them were not developed from statistical optimal signal detection perspective. We provide a method called TFisher, which is complementary to existing methods, and is shown more powerful in many cases. TFisher incorporates LD information in the testing process and it also simplifies computation to make it feasible for big data analysis, that is CCT based on the omnibus TFisher process. We applied the strategy to analyzing GEFOS data sets. The results show that TFisher has significantly higher positive rate among the group of known genes that we can found in literature so far. Some putative novel genes related to osteoporosis are reported that are shown relevant in data analysis of functionality and not detected by existing methods.