Application of full-genome analysis to diagnose rare monogenic disorders

Joseph T.c Shieh ,Hazel Perry,Daniah Beleford,Zhongxia Qi,Anne Slavotinek,David I K Martin ,Mónica Penón-Portmann ,Renata C Gallagher,Jingwei Yu,Andrew G Sharo,Bryce A Mendelsohn,Michal Levy‐Sakin ,Pui‐Yan Kwok ,Stephen Chow ,Ophir D Klein,Michelle Verghese,Jessica Tenney,Steven E Brenner,Karen Wong ,Dario Boffelli

doi:10.1038/s41525-021-00241-5

Abstract

Current genetic tests for rare diseases provide a diagnosis in only a modest proportion of cases. The Full-Genome Analysis method, FGA, combines long-range assembly and whole-genome sequencing to detect small variants, structural variants with breakpoint resolution, and phasing. We built a variant prioritization pipeline and tested FGA’s utility for diagnosis of rare diseases in a clinical setting. FGA identified structural variants and small variants with an overall diagnostic yield of 40% (20 of 50 cases) and 35% in exome-negative cases (8 of 23 cases), 4 of these were structural variants. FGA detected and mapped structural variants that are missed by short reads, including non-coding duplication, and phased variants across long distances of more than 180 kb. With the prioritization algorithm, longer DNA technologies could replace multiple tests for monogenic disorders and expand the range of variants detected. Our study suggests that genomes produced from technologies like FGA can improve variant detection and provide higher resolution genome maps for future application.

Highlights

Current approaches to diagnosis of monogenic conditions include short-read sequencing of exomes or genomes[1,2,3,4,5,6] the diagnostic yield from these methods is promising, ranging from 26 to 40%1,5, they leave many cases unresolved[2,7]
Using an automated genetic variant interpretation pipeline, we performed FGA on 50 undiagnosed cases to determine diagnostic yield and asked if it could help solve cases that had not been diagnosed with the previous testing
The automated pipeline integrates the longer DNA technologies into the diagnostic realm by enabling a streamlined variant detection protocol and minimizing biases introduced during the analysis process

Summary

Introduction

Current approaches to diagnosis of monogenic conditions include short-read sequencing of exomes or genomes[1,2,3,4,5,6] the diagnostic yield from these methods is promising, ranging from 26 to 40%1,5, they leave many cases unresolved[2,7]. Even whole-genome sequencing (WGS), which provides up to 9% additional diagnostic yield compared to exome sequencing[5,6,7,14] cannot detect all structural variants (especially duplications, inversions, and translocations), create chromosomal maps, or provide phasing information. Genetic diagnosis of rare disorders often entails “experiments of one,” where many sequence variants found in the proband must be vetted against current knowledge (gene/variants and genome reference) to decide if variants meet the diagnostic criteria[17]. Until we understand the functional consequences of more variants, or more patients with the same phenotypes are found, many candidate variants remain variants of uncertain significance

Methods

Results

Conclusion