Complex diseases are influenced by a multitude of genetic, environmental, and lifestyle factors, making causal inference and accurate risk prediction challenging. Our research aims to advance our understanding of complex diseases by integrating multi-modal data from genomics, phenomics, drug, and behavior and developing novel methods for causal inference and risk prediction.
1. Enhancing polygenic risk score (PRS) predictions: We will leverage cross-biobank studies to improve PRS predictions by incorporating biomarkers and disease networks. By accounting for complex interactions between genetic variants, molecular phenotypes, and clinical outcomes, we aim to develop more accurate and clinically actionable PRS.
2. Addressing horizontal pleiotropy in Mendelian randomization (MR): We will propose a novel MR framework that accounts for complex correlated horizontal pleiotropy and time-varying effects through disease-disease networks.
3. Phenome-wide association studies (PheWAS) and MR: We will conduct comprehensive PheWAS to identify novel causal associations between exposures and a wide range of traits. To address challenges in selecting appropriate instrumental variables in MR, we will explore various approaches, including data-driven methods and machine-learning techniques.
4. Disease-disease network: By integrating genetic data with phenotypic data, we aim to discover novel disease associations and develop disease networks.
