Description

Aldo Córdova-Palomera1, Jason Fries2,3, Paroma Varma4, Vincent S. Chen2, Madalina Fiterau2, Ke Xiao1, Heliodoro Tejeda1, Bernard Keavney5,6, Heather J. Cordell7, Christopher Ré2, Euan Ashley8, James R. Priest1

1Department of Pediatrics, Division of Pediatric Cardiology Stanford Medicine, Stanford, CA; 2Department of Computer Science, Stanford University, Stanford, CA; 3Center for Biomedical Informatics Research, Stanford University, Stanford, CA; 4Department of Electrical Engineering, Stanford University, Stanford, CA 5Cardiovascular Sciences, Faculty of Biology, Medicine & Health, University of Manchester, Manchester, UK; 6Manchester University NHS Foundation Trust, Manchester Academic Health Science Centre, Manchester, UK; 7Institute of Genetic Medicine, Newcastle University, Newcastle upon Tyne, UK; 8 Department of Medicine, Stanford University, Stanford, CA

With a prevalence of 1-2% in the general population, bicuspid aortic valve (BAV) is the most common congenital heart disease (CHD) and accounts for more morbidity and mortality than all other CHDs

combined. Although reported heritability estimates are as high as 89%, specific molecular genetic markers of BAV risk remain to be discovered.

Here, 9,802 magnetic resonance imaging (MRI) sequences from the UK Biobank were analyzed to classify aortic valves as either BAV or normal (tricuspid) using a deep learning algorithm (https://www.biorxiv.org/content/early/2018/06/05/339630). A genome-wide association study was conducted on the subset of unrelated European-ancestry participants (595 BAV, 9207 tricuspid aortic valve) using PLINK. External validation of the genetic findings was performed on imputed data from a case-control study of up to 2594 cases representing eight CHD types and 5159 healthy subjects from the Wellcome Trust Case Control Consortium 2 (WTCCC2).

Markers at three loci displayed statistically significant associations with BAV, including a variant on chromosome 12 near IGF1 and LINC00485 (rs146357447, 12:103025165, MAF=1.3%, odds ratio (OR): 3.2, p=6.1e-9), an intronic locus on MIR28 (rs550423221, 3:188508236, MAF=0.2%, OR=9.6) and a marker on chromosome 2 (rs192377594, 2:140363901, MAF=0.6%, OR=4.1). In the external dataset rs146357447 was associated with risk for atrial septal defect (OR=1.9, p=0.033), and the MIR28 marker displayed an association with non-specific/mixed CHD (OR=1.9, p=0.013).

The results suggest novel candidate loci as determinants of genetic risk for BAV in the general population, and indicate a shared genetic architecture with different types of CHD.