Inference of causal biomarker network to boost drug repositioning and minimise side effects
Epidemiological studies are predominantly observational, from which drawing causal inference remains a major challenge. Mendelian randomisation (MR) studies harness robust genetic associations to infer potential causal effect of an exposure on outcomes. In this project we will derive and integrate gene-disease and gene regulatory networks with data linking drugs and genes (Connectivity Map, DSigDB, etc.) to improve drug-repositioning efforts. All developed methods will be applied to simulated- and real data from large biobanks (UK, Estonia) with gene/protein expression levels as exposures and cardio-metabolic diseases as outcomes. The most promising therapeutic solutions will then be validated through a longitudinal study (FinnGen) with extensive information on medication use and health outcomes. The generated knowledge will accelerate drug discovery/repositioning.