Genetic architecture of cardiac dynamic flow volumes

Abstract Cardiac blood flow is a critical determinant of human health. However, definition of its genetic architecture is limited by the technical challenge of capturing dynamic flow volumes from cardiac imaging at scale. We present DeepFlow, a deep learning system to extract cardiac flow and volumes from phase contrast cardiac magnetic resonance imaging. A mixed linear model applied to 37,967 individuals from the UK Biobank reveals novel genome-wide significant associations across cardiac dynamic flow volumes including aortic forward velocity, total left ventricular stroke volume, forward left ventricular flow and aortic regurgitation fraction. Mendelian randomization using CAUSE reveals a causal role for aortic root size in aortic valve regurgitation. The most significant contributing variants (near ELN, FBN1 and ULK4) are implicated in connective tissue and blood pressure pathways. DeepFlow cardiac flow phenotyping at scale, combined with population-level genotyping data in the UK Biobank, reinforces the contribution of connective tissue genes, blood pressure and root size to aortic valve function in the general population.