Presented by: Dr. Jingjing Yang, Ph.D., Center for Computational and Quantitative Genetics, Assistant Professor of Human Genetics, Emory University School of Medicine

About the presentation: Existing transcriptome-wide association study (TWAS) tools require individual-level genetic and transcriptomic data from a reference source like GTEx (n=~100s) to estimate eQTL effect sizes. To extend the TWAS approach to leverage enormous summary-level molecular QTL data to map genetic risk genes, we develop a framework called OTTERS (Omnibus Transcriptome Test using Expression Reference Summary data) that can use summary-level eQTL datasets to perform TWAS. OTTERS adapts a variety of published polygenic risk score (PRS) methods (P+T, lassosum, SDPR, PRS-CS) to train eQTL effect sizes based on a multivariate regression model. For each PRS method, OTTERS uses the estimated eQTL weights to conduct TWAS, and then combines these TWAS p-values together to create an omnibus TWAS p-value per gene.Both simulation and real studies demonstrated that OTTERS could improve TWAS power by incorporating multiple statistical models to estimate eQTL effect sizes. We applied OTTERS to blood eQTL summary-level data (n=31,684) from the eQTLGen consortium and GWAS summary data of cardiovascular disease from the UK Biobank. OTTERS identified 38 independently significant risk genes, including 17 novel risk genes not identified by FUSION using GTEx reference panel.
Further, we showed that OTTERS can also be applied to integrate summary-level protein quantitative trait loci (pQTL) with GWAS data through the same framework. By applying OTTERS to summary-level pQTL data of brain, CSF, and plasma and GWAS data of Alzheimer’s disease (AD), we identified an interesting network of significant risk genes involving APOE, a well-known risk gene of AD, and immune function related genes. In conclusion, OTTERS not only provides a practical and powerful tool for TWAS analysis, but also provides the opportunity to leverage other emerging summary-level molecular QTL data, such as methylation, histone marks, and proteins. Free software implementing OTTERS is available on Github, https://github.com/daiqile96/OTTERS.

Join in-person in ASB 2200G, via Zoom or via the telephone.

To join via Zoom, visit https://pshealth.zoom.us/j/94860982030
To join via phone, dial: 929-205-6099
Meeting ID is 948 6098 2030 and the passcode to enter seminar is 320367.