Long-Read Sequencing in Genetic Medicine: Evaluating Haplotype Identification, Identity-by-Descent Analysis, and Diagnostic Yield

Abstract

Rare genetic diseases affect an estimated 400 million people worldwide, yet fewer than half will ever receive a molecular diagnosis. Current standard-of-care testing relies on short-read sequencing technologies that cannot span complex or repetitive genomic regions in a single read, leaving a substantial fraction of disease-causing variation undetected. A confirmed molecular diagnosis is essential for guiding clinical management, identifying potential therapies, informing reproductive decision-making, and ending the diagnostic odyssey. Long-read sequencing (LRS) addresses these limitations by generating reads that physically span repetitive elements, structural variants, and distant variant pairs, enabling accurate haplotype phasing — the determination of whether variants reside on the same or different chromosomes — without requiring parental samples. This dissertation evaluates three dimensions of LRS in genetic medicine: the technical determinants of haplotype phasing accuracy, the utility of LRS-based phasing for identity-by-descent detection, and the diagnostic yield and cost of LRS after non-diagnostic workup.