Expansion of Sequencing Technologies in Clinically Undiagnosed Genetic Disease

Abstract

The massive expansion of genomic technologies since the completion of the Human Genome Project has enabled the implementation of genetic sequencing into standard clinical care. Numerous common phenotypes, including cancer, developmental delay, and metabolic disease - as well as rare and undiagnosed conditions - are evaluated by gene panels and whole exome sequencing. Yet, despite this improved access to sequencing technology in the clinic, our ability to detect and interpret certain types of genetic variants remains severely limited. Whole exome sequencing, the most comprehensive test widely available for clinical use, only has a diagnostic rate of 30-50%, leaving the majority of patients undergoing genetic evaluation without a diagnosis. My dissertation work has aimed to improve this low rate of diagnosis in clinical evaluation through several methods. First, we have shown that rapid whole genome sequencing, an approach that takes advantage of machine learning to expedite variant discovery, has a high sensitivity to diagnose critical genetic conditions in the neonatal period when patients are too unstable for traditional sequencing turnaround time. Second, we have utilized RNA sequencing to detect regulatory aberrancies, such as mRNA splicing and changes in gene expression, that cannot be gleaned from analysis of the DNA sequence alone. This approach, combined with long-read DNA sequencing to detect novel structural events and non-coding variants, led to the detection of presumptive disease-causing variation in our cohort of undiagnosed patients. Lastly, we describe an approach to facilitate interpretation of variants of uncertain significance identified during newborn screening, to allow for early diagnosis and treatment initiation in patients. Together, the adaptation of these technologies increase the ability to detect and interpret genetic variants in patients currently undiagnosed from standard clinical sequencing methods.