Leveraging large-scale biobank data to investigate the landscape of congenital anomalies and to discover their genetic risk factors

Abstract

Congenital anomalies (CAs) affect 3-6% of live births and are the leading cause of infant morbidity and mortality. Large-scale genomic biobanks provide an opportunity to better understand the genetic drivers of CAs and reduce their public health burden. We utilized two electronic health record (EHR) linked biobanks, BioVU and the electronic medical records and genomics (eMERGE) consortium biobank, to investigate four CA populations: individuals with at least one CA, multiple congenital anomalies (MCA), congenital heart defects (CHDs), and craniofacial anomalies (CFAs). First, we developed a portable and accurate method for identifying individuals with MCAs in the EHR utilizing a novel approach removing distinctions between minor and major anomalies. We conducted genome wide association studies (GWAS) in all four CA populations. Initially, an analysis restricted to absorption, distribution, metabolism, and excretion (ADME) genes, which are crucial for teratogenic drug processing, revealed no significant SNP associations with CA, MCA, CFA, or CHD status. Expanding the analyses to all SNPs genome-wide identified serval significantly associated (p<1.0x10-6) SNPs within genes important for fetal development, such as PARD3B, involved in cell polarity. Transcriptome-wide association studies (TWAS) identified genes crucial for development whose genetically predicted gene expression (GPGE) were associated with CA status. The GPGE of NMNAT1, which is already associated in a Mendelian fashion with one MCA syndrome, was significantly associated with MCA status. Both eMERGE and BioVU CFA populations had a significant association with the GPGE of GLI2, which is involved with cardiomyogenesis. Our work demonstrates the value of utilizing big data resources for identifying genetic drivers of CAs and highlights the complex, shared nature of CAs. We plan to extend these analyses to additional biobanks.