Date/Time
Date(s) - 11/18/2024
3:00 pm - 4:00 pm
Location
Communicore, C1-15
Many of common human diseases are affected by genetic variations. Identifying risk genes of these diseases using genetic data have the potential of revealing molecular mechanisms of diseases and pointing to potential therapeutic targets. In this talk, I will present two methods for risk gene discovery. In the first part, I will focus on genomewide association studies (GWAS), the main approach for genetic study of common variants in the population. In particular, I will focus on a commonly used approach for risk gene discovery using expression quantitative trait loci (eQTL) data. This type of data reveals the genetic effect on gene expression. Existing methods for analyzing eQTLs, however, suffer from high false positive rates of risk gene discovery. Our method, causal-TWAS (cTWAS), borrows ideas from statistical fine-mapping to control for potential confounders that contribute to false positive results by existing methods. cTWAS showed calibrated false discovery rates in simulations, and its application on common traits discovered new candidate genes.
In the second part, I will talk about a method based on rare genetic variants. This type of variants have much lower frequencies in the populations, and are generally harder to study than common variants. Nevertheless, rare variants offer several advantages, e.g. they tend to have large effect sizes on disease risks. Many methods have been proposed for associating rare variants to phenotypes, particularly, the “gene burden” tests that aggregate information of all variants in a gene. These gene-based methods, however, often make unrealistic assumptions, and in practice, have low power of discovery. We developed a Bayesian method: MIxture model based Rare variant Analysis on GEnes (MIRAGE). MIRAGE captures the heterogeneity of variant effects and incorporates external information of potential functional effects of variants. We demonstrate in both simulations and analysis of an exome-sequencing dataset of Autism, that MIRAGE outperforms current methods for rare variant analysis. The top genes identified by MIRAGE are enriched with known or plausible Autism risk genes.
Bio:
Dr. Xin He is an Associate Professor at Department of Human Genetics, University of Chicago. Dr. He’s research focuses on developing computational methods to detect risk genes of complex dieases and to gain deeper insights into the disease mechanisms. His lab has published papers in leading journals such as Science, Nature Genetics and Nature Methods.