Analysis and interpretation of next-generation sequencing data for the identification of genetic variants involved in cardiovascular malformation

Abstract

Congenital cardiovascular malformation (CVM) affects 7/1000 live births. Approximately 20% of cases are caused by chromosomal and syndromic conditions. Rare Mendelian families segregating particular forms of CVM have also been described. Among the remaining 80% of non-syndromic cases, there is a familial predisposition implicating as yet unidentified genetic factors. Since the reproductive consequences to an individual of CVM are usually severe, evolutionary considerations suggest predisposing variants are likely to be rare. The overall aim of my PhD was to use next generation sequencing (NGS) methods to identify such rare, potentially disease causing variants in CVM. First, I developed a novel approach to calculate the sensitivity and specificity of NGS data in detecting variants using publicly available population frequency data. My aim was to provide a method that would yield sound estimates of the quality of a sequencing experiment without the need for additional genotyping in the sequenced samples. I developed such a method and demonstrated that it provided comparable results to methods using microarray data as a reference. Furthermore, I evaluated different variant calling pipelines and showed that they have a large effect on sensitivity and specificity. Following this, the NovoAlign-Samtools and BWA-Dindel pipelines were used to identify single base substitution and indel variants in three pedigrees, where predisposition to a different disease appears to segregate following an autosomal dominant mode of inheritance. I identified potentially causative variants segregating with disease in all three of the pedigrees. In the pedigrees with Dilated Cardiomyopathy and Hereditary Sclerosing Poikiloderma these variants were in plausible candidate genes. Finally, NGS was used to identify rare, potentially disease causing indel variants in patients with sporadic, non-syndromic forms of CVM characterised by chamber hypoplasia. Two indel calling pipelines were used as a means to increase confidence in the identified indels. These two pipelines achieved the highest sensitivity calls using the method described above. In the 133 cases, evaluated for 403 candidate genes, indels were identified in 4 known causative genes for human cardiovascular disease, namely MYL1, NOTCH1, TNNT2, and DSC2.