Relationships between Genotype and Phenotype in Cardiac Electromechanical Function

This project will examine relationships between genotype and electrical and mechanical phenotypes in a large cohort study: the UK Biobank. The goal is to determine how genetic and environmental factors influence electrical and mechanical features of the heart, and their associations with developing disease.

The first year will be split between constructing statistical shape models from the BioBank data to be used in heritability and GWAS studies as the phenotype, learning to apply machine learning methods to genetic, ECG and imaging data and performing a literature review focused on cardiac ECG structure and function relationships with genetic markers to form a solid basis for the project.

The second year will focus on ongoing ECG and structure/function analysis studies, addressing the predictive power of ECG scans in adverse cardiac function/shape features using deep learning and logistic regression/partial least squares regression methods. The aim is to produce a set of electro-mechanical features that are maximally correlated between ECG, shape and motion.

Years 3/4 will investigate correlations between genetic factors and electromechanical features, and then relate these to genotypic and environmental risk factors. Exome sequencing data will be available on a portion of the BioBank dataset, incorporating new genetic markers into the analysis. Genetic risk scores will be generated which it is hypothesised will associate strongly with adverse cardiac events.