Research

Our research aims to understand how disease-associated genetic variants interact with the rest of system, across layers of biological scales from molecular to disease traits, with particular focus on lung diseases. Toward this goal, we develop novel computational methods leveraging functional genomic data, network models, and large-scle biobanks.

Identifying genetic variants associated with lung disease

We develop computational tools to facilitate genetic mapping for both rare and common genetic disorders. Our current projects in this area are:

• developing a rare-variant association test accounting for the interaction between rare variants and the genetic background;
• developing a computational pipeline for Genome-Wide Association Studies (GWAS) using external shared controls;
• developing a pleiotropy-based genetic association test for underpowered genetic studies.

Mapping functional pathways which disease-associated genetic variants act through

Currently, human genetic studies are hindered by the difficulty to functionally interpret disease-associated genetic variants. To address this challenge, we are currently working on the following projects:

• decomposing GWAS signals into pathways by leveraging the cross-trait correlation structure;
• uncovering pathway-level organization of disease-associated regulatory variants from functional genomic data using a network model.

Predicting disease risks, subtypes, and progression of disease

Better understanding of disease-associated genetic variants and their function will ultimately lead to improved ability to predict disease outcomes. Our current projects in this area include:

• developing new polygenic risk score algorithms;
• organizing a community experiment to assess the accuracy of polygenic prediction models (CAGI PRS Challenge).

---

_{Powered by Jekyll and Poole}