Bayesian grouped regression with application to genome-wide association studies
Sparse Bayesian Grouped Regression with Application to Genome-Wide Association Studies
Supervisors names: Dr. Daniel F. Schmidt, Dr. Enes Makalic, Dr. Guoqi Qian, and Prof. John Hopper
Genome Wide Association Studies (GWAS) are a powerful tool for examining the relationship between genetic variants and a phenotype of interest, such as a particular disease. While a number of GWAS have led to the identification of single-nucleotide polymorphisms (SNPs) associated with various diseases, more associations are expected to be found. Current methodologies have some potential limitations due to the insufficient sample size and miss heritability. As a consequence, there is a need for improved and more sophisticated statistical methods, and penalised regression models, such as the Lasso, the Horseshoe, and Bayesian approaches, hold promise for this purpose.
The objective of my research project is to explore and develop modern statistical approaches for analysing GWAS data such as Bayesian grouped regression where we can fit data with grouped structures simultaneously and select important groups of variables that are associated with the disease. Bayesian grouped models allow gene and pathway-level analyses of GWAS association, and therefore, more informative associations are expected to be identified.
PhD scholarship and funding body: Graduate research studentship