A computational pipeline for identifying copy number variation from single nucleotide polymorphism data and applications to congenital heart disease

Abstract

Copy number variants (CNVs) are duplications or deletions of regions of the genome. CNVs, similar to single nucleotide variants (SNVs), range in frequency and severity in their effects on human disease. Despite the likely importance of CNVs in pathophysiology, comparatively fewer studies have examined the effect of CNVs on human disease as compared to SNVs. In this work, I first review the methods used for determination of CNVs from widely available SNV chip data and present data from a cohort of children with congenital heart disease (CHD) that finds that increased positive predictive value can be derived from looking at the overlap in CNVs called by two divergent methodologies (94.6%, 35/37). I then compare the prevalence of these validated CNVs and find that children with nonsyndromic CHD have a higher burden of large, gene-overlapping CNVs compared to controls (12.1% vs. 5.0%, P=0.00016). Moreover, through use of Cox proportional hazards regression, I present data that the presence of a CNV is associated with significantly decreased transplant-free survival after surgery (HR=3.42, 95% CI: 1.66-7.09, P=0.00090) with confounder adjustment. These data suggest CNVs can be determined with high accuracy and that CNV burden is an important modifier of survival after surgery for CHD.