Resolving Methodological Challenges in Genomics Research: Causality, Risk Prediction, and Reproducibility

University Of Michigan At Ann Arbor, Ann Arbor MI

Investigators

Linked publications, trials & patents

Abstract

Project Summary With the rapid advances in high-throughput sequencing technology, genetic and genomic studies have expanded significantly in scale and scope, enabling systematic surveys of large population cohorts and a wide spectrum of clinical and molecular phenotypes. Yet, the scientific community faces mounting challenges in translating data into meaningful insights for understanding disease origins and supporting therapeutic development. Our proposed research targets three inter- related methodological challenges in current genetics and genomics research: i) inference of causal molecular mechanisms underlying complex diseases using observational genetic data and genomic annotations; ii) improving risk prediction models using personalized genetic and genomic information; iii) statistical assessment of reproducibility patterns in genomic data, driven by either genuine biological heterogeneity or unwanted extraneous batch effects. We will develop statistically rigorous and computationally efficient computational methods and open-source software packages. Furthermore, we aim to advance computational techniques tailored for analyzing complex and large-scale genomic data.