Identification and Prioritization of Putative Pathogenic Structural Variants based on Functional Annotation

Abstract

Current genetic testing of patients performed to identify the molecular cause and potentially drive therapy decisions is predominately focused on small variation affecting coding regions due to the limitations of the underlying experimental methods. Long-read sequencing approaches have been shown to overcome these limitations allowing the detection of the entire spectrum of larger genomic alterations i.e. structural variants (SVs) with an unprecedented resolution potentially revealing previously undetected disease-causing mechanisms. In this thesis, we discuss the potential of long-read sequencing in combination with a functional annotation-based framework to identify non-coding pathogenic SVs in a cohort of limb malformation patients. In the process, we developed a pipeline that combines short- and long-read sequencing data, filters the detected SVs based on allele frequency, and applies an extensive functional annotation-based prioritization resulting in sets of candidate SVs for all involved patients. We also conduct a comprehensive comparison of callers and technologies highlighting the superior performance of long-read sequencing for SV detection and an evaluation of an automated prioritization method indicating superior performance to comparable approaches. The results of this thesis suggest the potential of performing an extended analysis of SVs as part of clinical diagnostics workflows and the relevance of non-coding functional annotation during variant prioritization.