A genome-wide approach for the discovery of novel repeat expansion disorders in the Undiagnosed Diseases Network cohort.

Abstract

Purpose The Undiagnosed Diseases Network (UDN) is a National Institutes of Health funded research study that aims to solve a broad clinical spectrum of challenging rare disease cases. Participants receive care from multiple clinical specialists, who collaborate to perform deep phenotyping and state-of-the-art multi-omics analyses. As bioinformatics of short-read sequencing has matured, the discovery of repeat expansion disorders (REDs) is accelerating. REDs comprise ∼60 characterized disorders, which exhibit a broad spectrum of phenotypes. Thus, a largely unbiased genome-wide approach in a phenotypically diverse sample will add to the diagnostic depth, explore the limits of short-read genome analysis, and establish novel candidate RED loci. Methods Here, we present a genome-wide analysis of repeat expansions conducted on 1,018 genomes from the UDN. By leveraging two distinct bioinformatics tools, ExpansionHunter Denovo and STRling, we show that repeat expansions can be accurately detected in short-read genomes. Results We demonstrate that a genotype-first approach can diagnose atypical cases of known REDs and provide valuable clinical insights. We present clinical details on participants with expansions in ATXN7, DMPK, FMR1, GLS, HTT, RFC1, AFF3, and MARCH6. Importantly, we highlight two cases of juvenile Huntington disease that were discovered through our analysis. Finally, we present a list of novel candidate TRs that could potentially be pathogenic if expanded. Conclusion Importantly, our approach showcases the bioinformatic advancements in genome analysis for RED detection and highlights its practical applications.