Skip to navigation Skip to main content Skip to footer

Approved research

Assesments of Joint Effect of Polygenic Disease-risk Scores and Modifiable Risk-factors on Major Chronic diseases and Mortality

Principal Investigator: Professor Nilanjan Chatterjee
Approved Research ID: 17712
Approval date: August 15th 2016

Lay summary

The aim of the research is to use data from UK Biobank prospective cohort study to estimate joint risk of multiple common disease conditions, including cancer, type-2 diabetes, cardiovascular diseases, and overall mortality, associated with GWAS generated polygenic disease-risk scores and modifiable risk-factors such as smoking, BMI, alcohol and physical activity. The analysis will allow understanding of how the potential impact intervention on modifiable risk-factors, may or may not vary by individual's genetic risk-profiles. The proposed research will generate valuable information regarding whether genetic information could be useful for targeting certain primary prevention efforts for risk-factor intervention that cannot be applied to the general population for cost and other burdens. For subjects in the UK Biobank cohort, polygenic risk-score (PRS) for major chronic diseases will be constructed based on published literature on susceptibility SNPs and their disease odds-ratios. Data from UK Biobank will be then used to estimate absolute risk of different disease endpoints and mortality in population strata defined by polygenic risk scores and modifiable risk-factors. For evaluating combined endpoint like overall mortality or overall cancer incidence, the disease-specific PRSs will be combined to form a composite PRS. Further, to understand the potential causal effect of intervention of modifiable risk-factors, like BMI, a Mendelian Randomization approach will be used to estimate absolute risk-reduction parameters associated with BMI for the different outcomes within strata defined by the disease associated PRS variables. These `instrumental` variable derived absolute-risk reduction parameters will then be then compared with more direct epidemiological estimate of the same parameters for the assessment of consistency of results across two types of analyses. All subjects in the full cohort with or without genotype data