Author:
Yu Mengyao,Harper Andrew R.,Aguirre Matthew,Pittman Maureen,Tcheandjieu Catherine,Amgalan Dulguun,Grace Christopher,Goel Anuj,Farrall Martin,Xiao Ke,Engreitz Jesse,Pollard Katherine,Watkins Hugh,Priest James R.
Abstract
ABSTRACTBackgroundThe interventricular septum (IVS) plays a primary role in cardiovascular physiology and a large proportion of genetic risk remains unexplained for structural heart disease involving the IVS such as hypertrophic cardiomyopathy (HCM) and ventricular septal defects (VSD).ObjectivesWe sought to develop a reproducible proxy of IVS structure from standard medical imaging, discover novel genetic determinants of IVS structure, and relate these loci to two rare diseases of the IVS.MethodsWe performed machine learning to estimate the cross-sectional area of the interventricular septum (IVS.csad) obtained from the 4-chamber view of cardiac MRI in 32,219 individuals from the UK Biobank. Using these extracted measurement of IVS.csad we performed phenome-wide association to relate this proxy measure to relevant clinical phenotypes, followed by genome-wide association studies and Mendelian Randomization.ResultsAutomated measures of IVS.csad were highly accurate, and strongly correlated with anthropometric measures, blood pressure, and diagnostic codes related to cardiovascular physiology. A Single nucleotide polymorphism in the intron of CDKN1A was associated with IVS.csad (rs2376620, Beta 8.4 mm2, 95% confidence intervals (CI) 5.8 to 11.0, p=2.0e-10), and a common inversion incorporating KANSL1 predicted to disrupt local chromatin structure was associated with an increase in IVS.csad (Beta 8.6 mm2, 95% CI 6.3-10.9, p=1.3e-13). Mendelian Randomization suggested that inheritance of a larger IVS.csad was causal for HCM (Beta 2.45 log odds ratio (OR) HCM per increase in SD of IVS.csad, standard error (SE) 0.48, pIVW = 2.8e-7) while inheritance of a smaller IVS.csad was causal for VSD (Beta −2.06 log odds ratio (OR) VSD per decrease in SD of IVS.csad, SE 0.75, pIVW = 0.006)ConclusionAutomated derivation of the cross sectional area of the IVS from the 4-chamber view allowed discovery of loci mapping to genes related to cardiac development and Mendelian disease. Inheritance of a genetic liability for either large or small interventricular septum, appears to confer risk for HCM or VSD respectively, which suggests that a considerable proportion of risk for structural and congenital heart disease may be localized to the common genetic determinants of cardiovascular anatomy.
Publisher
Cold Spring Harbor Laboratory