Harvard Catalyst Profiles

Polygenic scores are increasing in popularity as a way to use genetic variants throughout the genome to prioritize individuals at high risk of disease. However, they are not directly constructed to predict how people will respond to treatment (different strength of effect or side effects). We are working on implementing and understanding "interaction polygenic scores", designed specifically to address this differential response to environments, using a mix of computational approaches from statistical genetics and epidemiology. This opportunity for a student project would involve conducting genetic association studies in large cohort datasets like the UK Biobank, and using their results to compare approaches to developing these interaction polygenic scores. A specific area of clinical interest is the influence of supplements and body composition on cardiometabolic risk factors, but the project need not be limited to this area. Some computational experience (ex. R programming, Linux environments) is ideal but not required.