Targeted learning for causal population genomics

The Centre for Statistics awarded funding to an interdisciplinary team involving statisticians and the Institute for Genetics and Cancer.

Research Team: Sjoerd Beentjes, Ian Tomlinson, Andrew Bretherick, Chris Ponting and Dr. Ava Khamseh.

Project summary: Despite substantial investment in genome-wide association studies (GWAS) virtually all DNA variants that causally alter complex disease risk still elude us, stalling progress in our understanding of molecular disease mechanisms. Such causal genetic support is crucial for efficient drug discovery because it doubles the success rate of drugs in clinical development (King et al. 2019). Pinpointing trait-causal variants experimentally is painstaking and thus low-throughput, requiring cells that can only be assumed to be disease-relevant (Claussnitzer et al. 2020).

The aim of this project is to develop and apply state-of-the-art novel mathematical estimation methodologies in the semi-parametric estimation framework of Targeted Learning (TL), integrating causal machine learning and mathematical statistics, to answer such causal and statistical questions in population genomics and cancer.