Leveraging underrepresented population data improves interpretation of genetic variants associated with hearing loss

Abstract

Hearing loss is genetically heterogeneous, with over 121 implicated genes. Minor allele frequency (MAF) data from population databases greatly aid variant interpretation; however, these databases are predominantly based on individuals of European ancestry and lack sufficient East Asian representation, limiting accurate interpretation in under-represented populations. We analyzed rare variants associated with non-syndromic hearing loss classified as pathogenic, likely pathogenic, or of uncertain significance in the Deafness Variation Database (DVD). Population allele frequencies from 9,579 Koreans, 54,000 Japanese, and 651 patients with sensorineural hearing loss were evaluated. Of the 6,381 pathogenic or likely pathogenic variants cataloged in the DVD, 216 (3.38%) were detected in Korean population. Among these, 31 variants exhibited high allele frequencies that exceeded thresholds typically applied to identify benign variants in clinical interpretation guidelines. Of these, 6 remained disease-causing, including 4 East Asian founder alleles and one MYO7A variant common in Koreans. Our pipeline identified 24 variants for reclassification as benign or likely benign, and one P2RX2 variant of uncertain significance. Of 1,299,211 VUS, 3,736 were reclassified as benign. A substantial number of variants previously classified as pathogenic were reclassified as benign using MAF data from under-represented populations, highlighting the need for large-scale sequencing in diverse ancestries.