Identifying and understanding the function of trait associated non-coding variants (360G-Wellcome-105349_Z_14_Z)

In this project, I plan to expand on existing genome wide association study analysis techniques, to empower our ability to detect trait associated variants. The techniques will aim to target the ‘missing heritability’ and the current lack of knowledge surrounding the functionality of non-coding variants. This will be achieved by calculating, in an unbiased manner, the likelihoods of non-coding variants to control downstream gene products. In order to optimally evaluate the likelihood, I will impute publically available annotations into existing sizeable cohorts, for which whole-genome (or genotype) sequence data is available. These prior probability distributions will be subsequently incorporated into existing gene set test methods, in combination with multi-trait strategies. I will apply the developed method to existing cohorts, for which quantitative glycaemic and lipid trait measurements are available in an attempt to identify trait associated loci, which have not yet been uncovered. Additionally these likelihoods can be used to generate hypotheses, which explain how a variant mediates its effect on the observed phenotype. These theories will be subsequently tested, using methods that are able to elucidate causation from consequence and account for confounding factors.