Genetic landscape of copy number variations in congenital complex diseases

Abstract

Copy number variations (CNVs) are imbalanced structural variants characterized by alterations in the number of copies of DNA segments larger than 50bp. CNVs can disrupt coding sequences, alter gene dosage, and affect gene regulation. They are known to underlie Mendelian diseases and represent an essential portion of missing heritability in human complex diseases. It is therefore crucial to uncover the disease-associated CNVs and affected genes; however, CNV detection by current tools is prone to error and evaluating CNV accuracy remains challenging. Tetralogy of Fallot (TOF) and Hirschsprung disease (HSCR) are two rare congenital disorders, with the nonsyndromic form serving as models of complex diseases. Previous studies based on microarray or target sequencing technologies have discovered many CNVs associated with TOF or HSCR. Nevertheless, the genetic role of CNVs in these two complex diseases remains largely uncertain. To further explore the contribution of CNVs to TOF and HSCR susceptibility, especially those encompassing non-coding regulatory regions and those with small size (≤1kb) that were largely underestimated in previous studies due to technical limitations, two CNV studies were performed based on the whole genome sequencing (WGS) data of 150 nonsyndromic TOF case-parent trios, as well as the WGS data of 443 short-segment HSCR patients and 493 ethnically matched controls, respectively. CNVs were detected based on a newly developed framework, CNV-JACG, which was created for Judging the Accuracy of CNVs and Genotyping using paired-end WGS data. It is based on a random forest model trained on 21 distinctive features characterizing the CNV region and its breakpoints. CNV-JACG was demonstrated to have superior sensitivity and precision compared to the latest method, particularly for the small CNVs (≤1kb). The de novo CNV rate (including mosaic) in this TOF cohort was comparable with previous reports. In particular, a few candidate genes (HAND2, SORBS2, SMCHD1, MYOM1, PTPRM) were discovered through the integration of epigenomic and expression data. Among the rare inherited CNVs, compound heterozygous variants consisting of rare inherited deletions and missense point mutations were also identified wherein the well-documented congenital heart disease associated gene DNAH11 was involved. Four large-scale rare CNVs were uniquely identified in HSCR patients, three of them disrupted known enteric nervous system associated genes, EDNRB, SPRY2 and UCHL1. The category-wide association study found that duplications encompassing the enhancer/H3K27ac regions of the neural crest cell expressed/constrained genes were enriched in HSCR patients. Through new tool development and CNV analysis of both family and population WGS data, this thesis presents a useful tool for assessing the accuracy of CNVs to meet the ever-growing needs for uncovering the missing heritability linked to CNVs. The CNVs and affected genes discovered in this study extend the current understanding of the contribution of CNVs to TOF and HSCR susceptibility, with the potential of informing therapeutic decision-making.