Abstract
Purpose :
Improving next-generation sequencing (NGS) diagnostic accuracy is vital, requiring careful mitigation of false negatives. Common filters like gnomAD variant allele frequencies (AF) aim to exclude variants too common for rare diseases. Intrigued by the risk factor c.5603A>T p.(N1868I) in ABCA4 with gnomAD popmax AF 0.067 (v2.1.1), we hypothesize that numerous relatively common pathogenic variants may be misclassified as variants of uncertain significance (VUS) or benign, resulting in false-negative molecular diagnoses. Our goal is to establish a common pathogenic variant atlas for inherited retinal diseases (IRDs).
Methods :
In the pilot phase, a literature review identified prevalent IRD genes (ABCA4, USH2A, EYS). Variants from these genes listed as DM/DFP in HGMD professional version 2023.3 were processed through an in-house variant annotation pipeline. GnomAD v2.1.1 or v3.1.2 popmax AF ranging from 0.005 to 0.1 was considered. Variants with benign or conflicting interpretations in ClinVar were classified by ACMG/AMP 2015 guidelines.
Results :
Fourteen ABCA4, seven EYS, and eleven USH2A variants were listed as DM/DFP in HGMD, were listed as benign or had conflicting interpretations in ClinVar, and had a relatively common gnomAD v3.1.2 popmax AF (0.003 to 0.067). Variant curation reclassified 7/14 ABCA4 variants, classifying three pathogenic, two likely pathogenic, and two risk-factor variants; the remaining seven were VUS. In USH2A, one was classified as likely pathogenic, five as VUS, and four as likely benign. In EYS, one was classified as likely pathogenic, five as VUS, and one as likely benign. Updated classifications prompted genetic testing report amendments, providing molecular diagnoses for three patients.
Conclusions :
Our pilot study demonstrates that misclassified common pathogenic variants represent a significant source of false-negative findings in molecular diagnosis of IRDs. Our ongoing efforts aim to create a complete atlas for common pathogenic variants in all known IRD genes to enhance NGS diagnostic utility and validity.
This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.