AbstractHearing loss is a clinically and genetically heterogeneous sensorineural disease that affects approximately 1 out of 1000 newborns. For the molecular diagnosis of genetic hearing loss, target panel or whole-exome sequencing (WES) have been widely used due to their cost-effectiveness and efficacy. Despite the advantages of WES, the plausible diagnoses in a substantial number of patients remain elusive due to its limited coverage. Here we utilized whole-genome sequencing (WGS) on a large cohort of individuals with hearing loss to overcome the drawbacks of WES and find the advantages of WGS. We implemented a systematic workflow to identify coding region variants, cryptic splice variants, mitochondrial variants, copy number variants, cis-regulatory variants and transposable element insertions. WGS was conducted on 140 families with hearing loss. Causative variations were identified in 37 of these families, accounting for 26% of the total. WGS possessed the capability to find genetic variations that are not identifiable using WES. The identified variants by WGS in this study encompassed aberrant splicing variants in EYA1 and CDH23, mitochondrial variants in MT-RNR1 and MT-CO1, structural variants in STRC, and Alu insertion in SLC17A8. These findings highlight the benefits of WGS. With the decreasing cost of WGS, its usage will become more prevalent, allowing more precise identification of the genetic causes of hearing loss.
IntroductionHearing loss is the most common sensory disorder, affecting 1.33 per 1,000 newborns in developed countries1 and about half of individuals aged over 60 years (ref.2). It has a substantial impact on an individual’s quality of life. Accurate molecular diagnosis is crucial for identiyfing the genetic causes of hearing loss, allowing clinicians to forecast patient outcomes and select suitable therapeutic interventions3. However, achieving accurate genetic diagnoses in patients with hearing loss can be challenging due to the genetic heterogeneity and variable phenotypic expressivity of the condition4.Whole-exome sequencing (WES) has been widely used for genetic diagnosis in patients with hearing loss. WES primarily targets the exome, which represents only a small portion, approximately 1–2%, of the entire genome. Regarding hearing loss, numerous efforts for deciphering genetic etiologies underlying hearing loss have been made in different ethnicities by WES and targeted exome sequencing. For example, Guan et al. developed a tiered exome-sequencing-based comprehensive gene panel, named AUDIOME, for the diagnosis of heterogeneous nonsyndromic sensorineural hearing loss (SNHL), and screening resulted in a diagnostic yield of 33.3% (11/33)5. Shearer et al. developed a platform, OtoSCOPE, that couples targeted genomic enrichment and massively parallel sequencing to capture and sequence exon regions of genes associated with nonsyndromic hearing loss, and the overall diagnostic rate was 42% (42/100)6. In Australia, 56% (59/106) of the molecular causes of congenital hearing impairment in a population-derived 2-year birth cohort of infants were diagnosed by WES7. Besides, targeted genome sequencing of 1,027 Chinese patients with bilateral hearing loss revealed 57.3% (558/1027) of genetic causes8, and clinical exome sequencing of 47 English patients with SNHL resulted in diagnostic yield of 38.3% (18/47)9. In France, a combined multistep strategy including DFNB1 locus analysis, targeted massive parallel exome sequencing of 74 genes, and additional approaches including copy number variation, in silico analyses and minigene studies was conducted among French families with deafness, and causative variants were discovered in 47.8% of families (99/207)10. In the Republic of Korea, many studies reported the identification of pathogenic variants by WES in patients with deafness, and functional validation of those variants was performed11,12,13,14,15,16.Even though WES and clinical panel screening substantially contributed to the diagnosis of genetic hearing loss and has provided valuable insights into the genetic basis of hearing loss, WES techniques still have many limitations. For instance, WES cannot include regulatory or intronic regions that are known to be associated with a specific phenotype17. Furthermore, the ability to detect copy number variants (CNVs), especially those affecting one or a few exons, is so limited that important structural variants may be overlooked17. When multiple types of exome are available, WES is limited by the fact that not all tests are equivalent17.In recent years, whole-genome sequencing (WGS) has emerged as a powerful tool for genetic analysis. Indeed, WGS is considered an initial test for suspected mendelian conditions and also appropriate when no candidate variants are detected by WES17. In contrast to WES, WGS covers the entire genome, offering a more thorough perspective of the genetic landscape. This broader coverage of the genome enables the detection of variants in cis-regulatory regions, deep-intronic splicing alterations, mitochondrial DNA (mtDNA) variants and structural variants, which are increasingly recognized as critical contributors to gene expression and disease pathogenesis18,19,20.In this study, we sought to evaluate the diagnostic capabilities of WGS for patients with hearing loss. We hypothesized that WGS would provide additional insights into the genetic etiology of hearing loss by uncovering putative disease-causing variants that are difficult to be detected by WES. Here, we present the diagnostic yield of WGS and discuss the clinical utility of WGS in improving molecular diagnoses, predicting patient prognosis and guiding personalized treatment strategies.Materials and methodsStudy design and patientsThis study received approval from the institutional review board of Severance Hospital, Yonsei University Health System (IRB #4-2015-0659). Individuals with hearing loss were enrolled in the Yonsei University Hearing Loss (YUHL) cohort, and they provided informed consent for the participation in the study and publication of their clinical data. Among the total 729 enrolled families, WES was performed on probands from 437 families, while probands from 140 families underwent WGS.Clinical evaluationAll enrolled patients underwent a comprehensive physical examination and history interviews. Audiological evaluations were conducted for all patients, as well as their affected and unaffected family members, using appropriate tests such as pure-tone audiometry, otoacoustic emission test, auditory brain stem response test or auditory steady-state response test. For pure-tone audiometry, air and bone conduction thresholds were measured in a double-walled audio booth, with frequencies ranging from 250 to 8000 Hz for air conduction and 250 to 4000 Hz for bone conduction. The degree of hearing loss was categorized as mild (26–40 dB), moderate (41–70 dB), severe (71–90 dB) or profound (>90 dB), based on the average threshold of the four frequencies (500, 1000, 2000 and 4000 Hz). The audiogram pattern was defined as ascending when the average thresholds for high frequencies (2000 Hz and 4,000 Hz) were at least 25 dB lower than those for low frequencies (250 and 500 Hz), ski-sloping when high frequencies were more than 25 dB higher than low frequencies, and flat when the difference between high and low frequencies was within 25 dB.Variant callingGenomic DNA was extracted from peripheral blood obtained from the affected individuals and their parents (whenever available) using red blood cell lysis, cell lysis and protein precipitation solutions (iNtRon Biotechnology). The integrity of the genomic DNA was checked by running an agarose gel electrophoresis, and genomic DNA was quantified using fluorometry (Qubit, Invitrogen). The sequencing libraries were prepared according to the manufacturer’s instructions of TruSeq DNA Nano Library Prep Kit. (Illumina). In brief, 100 ng of genomic DNA from each sample were fragmented by acoustic shearing on a Qsonica 800 R2 instrument. Fragments of 350 bp were ligated to Illumina’s adapters and polymerase chain reaction (PCR)-amplified. An appropriate size for the final library is 500–600 bp. Libraries were quantified using the TapeStation 4200 instrument (Agilent Technologies) and KAPA Library Quantification Kit (KK4824, Kapa Biosystems). The resulting purified libraries were applied to an Illumina flow cell for cluster generation and sequenced using 150-bp paired-end reads on an Illumina NovaSeq 6000 (Illumina) sequencer by following the manufacturer’s protocols. The average sequencing depth of the libraries was 30×. Genomic Analysis Toolkit (GATK) best-practice pipelines were used to generate a binary alignment map (BAM) and variant call format (VCF) files from raw unmapped reads. The human reference genome GRCh38/hg38 was used to align the reads using the Burrows–Wheeler Aligner (BWA-MEM) algorithm. HaplotypeCaller was used to generate genotype VCF (gVCF) files for each sample. To filter out low-quality single-nucleotide variants (SNVs), a filter was applied with a coverage depth (CD) ≥5 and genotype quality (GQ) ≥20. To screen cytomegalovirus (CMV) infection among WGS samples, reads that were not mapped to the human reference genome were extracted using samtools (v1.3.0). Later, these unmapped reads were converted to the raw sequencing reads using bamToFastq of bedtools (v2.30.0), and paired fastq was realigned to the CMV reference genome (NCBI Accession NC_006273.2) with the BWA-MEM algorithm. The number and proportion of mapped reads to viral genome were calculated using flagstat function of samtools (v1.3.0). For mitochondrial short-variant discovery, GATK best-practice pipelines were used. Unlike germline-derived variants, mitochondrial variants were called using Mutect2. Blacklisted regions in mitochondrial genome were filtered, and variants having a ‘PASS’ flag in the ‘FILTER’ field were used for downstream analysis to remove false-positive calls. For CNV calling, a read-depth-based algorithm, CNVnator (v0.3.2, https://github.com/abyzovlab/CNVnator), was used as previously reported. For calling of transposable element (TE) insertion (TEI) from WGS data, computational tool xTea (v0.1.9, https://github.com/parklab/xTea) was used as reported.Coverage calculationCoverage was calculated on 19 probands who had undergone both WES and WGS for precise comparison. The depth of sequencing coverage was calculated using the tool Mosdepth (v0.3.3, https://github.com/brentp/mosdepth). For WES samples, an Agilent SureSelect V5 enrichment capture bed was given as interval options and the depth of corresponding regions was calculated. Gene coverage was calculated using DepthOfCoverage from GATK. First, among known hearing loss genes, only regions overlapping with an Agilent SureSelect V5 enrichment capture were obtained by bedtools (v2.30.0, https://github.com/arq5x/bedtools). We divided the percentage of a given region covered above 15 into 10 bins and counted the number of intervals in each sample that corresponded to each percentage bin. We then averaged the number of those intervals across samples, converted them to percentages and plotted them as a pie chart. Three main quality parameters, CD, GQ and minor read ratio (MRR) were calculated as previously reported21 and source code used for this analysis is publicly available via GitHub (https://github.com/HGID/WES_vs_WGS).Annotation and classification of candidate variants in known NSHL genesThe SNVs and small indels were annotated using ANNOVAR software. To identify rare variants of unknown significance, variants were filtered using total minor allele frequency (MAF) in the Genome Aggregation Database (gnomAD) with a cutoff of 0.005. Ethnicity-specific MAFs and POPMAX 95% confidence interval estimates were compared within gnomAD. Novel variants that are not reported in the genome aggregation database were also selected. The variants were prioritized using various in silico prediction scores. The pathogenicity of missense variants was predicted using prediction scores from at least five prediction tools: SIFT (https://sift.bii.a-star.edu.sg/), PolyPhen2 (http://genetics.bwh.harvard.edu/pph2/), MutationTaster2 (http://www.mutationtaster.org/), CADD (https://cadd.gs.washington.edu/) and REVEL (https://sites.google.com/site/revelgenomics/). The interpretation of variants was based on the clinical interpretation of the ClinVar (https://www.ncbi.nlm.nih.gov/clinvar/), Human Gene Mutation Database (http://www.hgmd.cf.ac.uk/ac/index.php) and Deafness Variation Database (https://deafnessvariationdatabase.org/) to identify previously reported pathogenic or likely pathogenic variants. If no interpretation was registered, the identified variants were classified as variants of unknown significance. Cryptic splice site variants were identified using SpliceAI as a plugin via Ensembl’s Variant Effect Predictor tool (v106, https://github.com/Ensembl/ensembl-vep) and filtered by a delta score threshold of 0.1 in transcribed regions of genes. For CNV, identified regions were annotated using a AnnotSV (v2.2, https://github.com/lgmgeo/AnnotSV). To increase confidence, only CNV regions >1 kb in size and variants with a ranking score above 3 from AnnotSV were selected. All candidate CNVs were manually confirmed via genome-wide visualization against control samples utilizing Integrative genomics viewer (IGV) (v2.13.2, https://github.com/igvteam/igv). Mitochondrial variants were additionally annotated using MITOMAP database (https://www.mitomap.org/MITOMAP) and prioritized by a disease status of ‘Cfrm’ or ‘Reported’ for deafness. Mitochondrial variants having GenBank or gnomAD MAF under 0.005 and representing maternal inheritance by segregation analysis were selected as causative variants. Among causative variants, variants that were previously reported as pathogenic when considering the ethnicity were regarded as highly pathogenic variants. As for TEIs, only insertions predicted by xTea as ‘two-sided target-primed reverse transcription’ were considered highly confident TEIs. These candidates manifest both breakpoints and have reads supporting a target site duplication and a poly-A tail. In addition, TEIs found in two or fewer patients in cohort and showing an MAF under 0.005 or not reported in the gnomAD-SV database were elected as highly presumable insertion candidates. Variants that are in 926,545 of human candidate cis-regulatory elements (cCREs) and hearing-loss-associated genes were annotated by retrieving data from the Search Candidate cis-Regulatory Elements by ENCODE (SCREEN) database of phase III of the Encyclopedia of DNA Elements (ENCODE) Project22. An MAF cutoff of 0.005 and internal frequency were also applied to the selection of cis-regulatory element variants as in TEI filtration. In silico prediction of gene perturbation was confirmed using the R package motifbreakR. Throughout the process of selecting candidate variants, the interpretation of clinicians and the inheritance pattern of probands were included. All candidate genetic variants were further confirmed by segregation analysis using Sanger sequencing, PCR or Multiple ligation-dependent probe amplification (MLPA).ResultsWGS demonstrated a diagnostic yield comparable to to that of WESFor the analysis of WGS, we established a multistep analysis process encompassing the detection of coding region variants, cryptic splice variants, cis-regulatory element variants, CNVs, mitochondrial variants and TEI (Fig. 1). For the 19 patient samples that underwent both WES and WGS, we conducted a comparison of sequencing depth, gene coverage and variant quality (Supplementary Figs. 1–3). WES demonstrated an average depth of 63× of the kit capture regions, and WGS had an average depth of 37× of the total regions (Supplementary Fig. 1a). For previously reported 173 genes associated with hearing loss (https://hereditaryhearingloss.org/), WES demonstrated an average depth of 53× and WGS had an average depth of 42× (Supplementary Fig. 1b). These targeted hearing-loss-associated loci were covered similarly in both sequencing methods, which was 99.6% (3467/3482) in WES and 99.1% (3449/3482) in WGS (Supplementary Fig. 2). In the WES analysis, about 72% of intervals had 90% of bases or more with depth above 15 (Supplementary Fig. 2a). Meanwhile, in the case of WGS, about 95% of intervals had 90% of bases or more with depth above 15 (Supplementary Fig. 2b). In particular, compared with WGS, a higher proportion of intervals had 60% of bases or less with depth above 15 in WES (Supplementary Fig. 2a, b). Notably, loci demonstrating coverage of 0% were limited to a few exonic regions of BDP1 in WES and primarily COL11A2 in WES and WGS. Poor coverage of several regions in BDP1 and COL11A2 may be attributed to the presence of adjacent alternative contigs in chromosomes 5 and 6. To assess the quality of SNV called by GATK practices, three main quality parameters for variants, CD, GQ and MRR, retrieved using WGS and WES in hearing-loss-associated loci were compared (Supplementary Fig. 3a–c). Overall, variants called by WGS showed higher and more consistent quality than those called by WES (Supplementary Fig. 3).Fig. 1: Workflow of WGS analysis.A flowchart illustrating the analysis process of WGS. Coding region variants, splice variants, CNVs, variants in the cCREs, mitochondrial variants, TEI, CMV infection and American College of Medical Genetics and Genomics secondary finding (ACMG SF) variants were detected and analyzed. DVD, Deafness Variation Database.Full size imageBefore conducting genetic testing on patients with hearing loss, we performed a prescreening for pathogenic variants in GJB2 and SLC26A4, the two most prevalent causative genes for nonsyndromic hearing loss in Korean individuals23,24. Among the 729 families enrolled in the YUHL cohort, 19% (141/729) of the families harbored pathogenic variants in SLC26A4 and 4% (30/729) had variants in GJB2 (Fig. 2a). Out of the 729 families, 437 were referred to WES, while 140 families underwent WGS (Fig. 2a). As mentioned earlier, 19 individuals underwent both WES and WGS. In the case of WES, diagnostic yield was approximately 35% (151/437) of cases (Fig. 2b). The frequent causative genes for families exhibiting an autosomal recessive inheritance pattern were MYO15A (10/437), CDH23 (8/437) and MPZL2 (6/437). Meanwhile, KCNQ4 (14/437), COCH (7/437), MYH14 (7/437) and MYO7A (7/437) were common causes of families showing an autosomal dominant inheritance pattern (Fig. 2b). The diagnostic yield was similar regardless of the presence of vertigo or the age of onset. However, patients with a family history of multiple affected individuals, autosomal dominant inheritance or severe or syndromic hearing loss demonstrated a higher rate of successful diagnoses (Fig. 2b).Fig. 2: Comprehensive WGS analyses exhibited diagnostic rates comparable to those with WGS for genetic hearing loss.a A pie chart representing the distribution of patients subjected to genetic screening in a cohort study. Before next-generation sequencing, patients were prescreened for variants in GJB2 and SLC26A4, the most common genes responsible for hearing loss among the Korean population. b, c The diagnostic rate and gene composition of patients screened by WES (b) and WGS (c). Red indicates genes showing an autosomal recessive inheritance pattern, and yellow indicates genes showing an autosomal dominant inheritance pattern. Proportions of diagnosed patients corresponding to clinical manifestations, including vertigo, age of onset, mode of inheritance, syndromic feature and severity of hearing loss, are represented as bar plots; blue, diagnosed; light gray, undiagnosed.Full size imageThe diagnostic rate of WGS was approximately 26% (37/140) (Fig. 2c). The diagnostic rate seen in our cohort was lower for WGS compared with WES. However, it was challenging to directly compare the diagnostic rates between WES and WGS owing to considerable differences in the patient characteristics recruited for each sequencing method (Supplementary Table 1). The most frequent causative gene for families exhibiting an autosomal recessive inheritance pattern was MYO15A (4/140). Meanwhile, CHD7 (3/140), TECTA (2/140), KCNQ4 (2/140) and POU4F3 (2/140) were detected in families showing an autosomal dominant inheritance pattern (Fig. 2c and Supplementary Table 2). Interestingly, some patients who underwent WGS were identified to have pathogenic CNVs (2/140), a cryptic splice variant (2/140) and mtDNA variants (4/140), which are difficult to detect with WES. Patients who exhibited postlingual, autosomal recessive inheritance, severe hearing loss or syndromic hearing loss rather than nonsyndromic hearing loss showed higher diagnostic rates with WGS (Fig. 2c).Aberrant splice variants located outside the core splice sites were identified through WGSFor patients referred to WGS, a comprehensive analysis of splicing variant was conducted because WGS covers a broader intergenic and intronic region of the genome, rather than being restricted to exon-flanking intronic regions25. SpliceAI, a 32-layer deep neural network that predicts splicing from a pre-mRNA sequence, was utilized for the detection of potential cryptic splice variants26. We filtered and prioritized the detected splice variants that had a delta score higher than 0.1, indicating a high probability of splice alteration (Supplementary Table 3).In the YUHL121 family, we identified the c.640-6T>A variant in the eighth intron of EYA1 in patient YUHL121-21 through WGS (Fig. 3a). The region harboring the variant was covered by both WES and WGS, although the patient was screened only by WGS. Variants in EYA1 are the most common cause of branchio-oto-renal syndrome, typically inherited in an autosomal dominant manner27. The proband (YUHL121-21) was a 10-month-old male referred for bilateral congenital moderate hearing loss. The click Auditory brainstem response (ABR) test showed thresholds of 60 dB for both ears, and the distortion-product otoacoustic emission (DPOAE) test was negative for both ears (Fig. 3b). He had a family history of hearing loss in his father, who exhibited mild conductive hearing loss in mid-to-low frequencies (Fig. 3b). Clinical examination revealed bilateral preauricular pits without other external ear anomalies. The proband had previously undergone excisional surgery for bilateral second branchial cleft cyst at another hospital 3 months before visiting our clinics. Genitourinary ultrasonography showed no evidence of renal anomalies. Based on these clinical features, the clinician suspected branchio-oto-renal syndrome according to the diagnostic criteria proposed by Chang et al.28. Segregation analysis by Sanger sequencing in the YUHL121 family verified the variant and revealed that the heterozygous variant was inherited from the father (Fig. 3c). Furthermore, the detected variant was never reported in the population, according to the Genome Aggregation Database (gnomAD), and predicted to cause aberrant mRNA splicing, creating a premature terminating codon and truncation of the C-terminal EYA domain, where most of the reported variants are clustered (Fig. 3a)29. In addition, the probability of being loss-of-function intolerant (pLI) score of the EYA1 gene is 1.0 according to the gnomAD, indicating that this gene is highly likely to be haploinsufficient. In summary, co-segregation within an affected family member and the molecular characteristics, combined with evident clinical phenotypes, led us to classify the c.640-6T>A variant as potentially pathogenic.Fig. 3: Deep-learning-based splicing prediction of WGS identified cryptic splice mutations outside of core splice site.a A pedigree of proband YUHL121-21 and graphical scheme of predicted gene perturbation by aberrant splicing variant, c.640-6T>A of EYA1. The altered allele is labeled in red, and the predicted probability of variant acting as a splice acceptor is indicated; red hexagon, premature stop codon. b Distortion-product otoacoustic emission (DPOAE) (left) and pure-tone audiometry (PTA) (right) results of YUHL121-21 and YUHL121-11, respectively. For DPOAE, dashed lines indicate the noise floor, and solid lines indicate detected distortion products. AC, air conduction thresholds; BC, bone conduction thresholds. c Segregation results of a c.640-6T>A variant in YUHL121 family. Sanger sequencing chromatograms revealed that c.640-6T>A from was inherited from the father of a proband. -21, proband; -11, father; -12, mother. d A pedigree of the YUHL704 family and a graphical scheme of the predicted gene perturbations by aberrant splicing variants in the CDH23 gene. The altered allele is labeled in red, and the predicted probability of variants acting as cryptic splice acceptor or donor is indicated; red hexagon, premature stop codon. e Segregation results of the c.6050-9G>A and the c.6050-173G>A variants in the YUHL704 family. -21, proband; -12, mother. f PTA results of YUHL704-21. AC, air conduction thresholds; BC, bone conduction thresholds.Full size imageIn the YUHL704 family, we identified compound heterozygous variants of the CDH23 gene, the c.6050-9G>A near the 47th exon and a deep-intronic c.6050-173G>A variant between the 46th and 47th exons through WGS (Fig. 3d). The c.6050-9G>A variant was in a region covered by both WES and WGS, whereas the c.6050-173G>A variant showed poor coverage or no coverage at all by WES. The c.6050-9G>A variant was previously reported as pathogenic in several studies30,31. This variant is predicted to create a cryptic splicing acceptor, inserting seven additional nucleotides from the 46th intron and introducing a premature stop codon in the following exon, resulting in nonsense-mediated decay of the transcript30. Furthermore, the c.6050-173G>A variant in the deep-intronic region of 46th intron is predicted to create a cryptic splicing donor, which might induce the extension of the 46th exon and the introduction of a premature termination codon. Moreover, this variant has never been reported in gnomAD. Segregation analysis by Sanger sequencing of the YUHL704 family revealed that the c.6050-9G>A variant was inherited from the mother, while the c.6050-173G>A variant was not detected in her, indicating that the two variants are in a trans state (Fig. 3e). The proband was a 63-year-old male and exhibited progressive hearing loss initiated in his 20s without definite vestibular dysfunction. Audiometric evaluation indicated bilateral, moderate, ski-sloping hearing loss (Fig. 3f). These clinical characteristics are consistent with DFNB12, caused by variants in the CDH23 gene and presenting with a milder phenotype compared with type 1D Usher syndrome32. Therefore, we concluded that the identified variants potentially explain his hearing loss, which would otherwise be unexplainable by WES owing to the hidden deep-intronic variant.WGS detected pathogenic mtDNA mutationsTo eliminate sequencer noise and contamination of inserted mtDNA into the nuclear genome or noncoding control region, we followed the best practices of the mitochondrial short-variant discovery pipeline of GATK (Fig. 1) for the discovery of variants in mitochondria DNA. Due to the limited capacity of WES in detecting mitochondrial variants19, only WGS samples were used for variant discovery. Initially, a total of 1338 mtDNA variants were identified through the analysis. Among them, allele frequencies for 119 mtDNA variants were available from GenBank and a few variants previously reported to be associated with hearing loss were detected in our cohort (Fig. 4a and Supplementary Table 4). Upon thorough review of the clinical history and inspection of copies of detected mtDNA variants, hearing loss in a patient (YUHL58-21) carrying the homoplasmic m.7444G>A mutation in MT-CO1, and in three patients (YUHL390-21, YUHL681-21 and YUHL847-21) harboring the homoplasmic m.1555A>G mutation in MT-RNR1, was determined to be associated with mtDNA variants (Fig. 4b, c).Fig. 4: Pathogenic homoplasmic m.1555A>G and m.7444G>A were responsible for maternally inherited hearing loss.a A Circos plot of the mitochondrial genome. The scatter plot and heat map represent the retrieved GenBank frequency of 119 variants among total detected mitochondrial variants in patients screened by WGS. Putative pathogenic variants, m.1555A>G and m.7444G>A, are labeled in red. The histogram represents the number of patients harboring variants present in a region when the chromosome is divided into 500-bp bins. b A pedigree of proband YUHL58-21 and segregation results of a m.7444G>A variant in the YUHL58 family. Sanger sequencing chromatograms revealed that homoplasmic m.7444G>A was maternally inherited. -21, proband; -11, father; -12, mother; -22, sibling; ctrl, control. c A pedigree of proband YUHL390-21, YUHL681-21 and YUHL847-21 and segregation results of a m.1555A>G variant in the corresponding family. Detected homoplasmic m.1555A>G are further confirmed by Sanger sequencing. -21, proband; -22, sibling.Full size imageThe m.7444G>A variant has been causally linked to deafness in several studies and has also been assumed to influence the penetrance and expressivity of the m.1555A>G mutation33,34. In addition, the m.1555A>G mutation is commonly reported to be associated with nonsyndromic and aminoglycoside-induced hearing loss35. The clinical characteristics of the four patients identified with the m.1555A>G and m.7444G>A mutations are summarized in Supplementary Table 5. Both the m.7444G>A and m.1555A>G mutations lead to hearing loss with variable expressivity36,37. Consistently, the four patients exhibited heterogeneous phenotypes, ranging from bilateral, congenital and profound hearing loss to asymmetric, postlingual or mild hearing loss. None of the patients reported exposure to aminoglycosides, although an accurate record of administration was impossible due to the retrospective nature of the history interview. The hearing loss in YUHL390, YUHL681 and YUHL847 was familial, and that in YUHL681 and YUHL847 was suspected to be maternally transmitted, although incomplete penetrance was observed. Interestingly, the mother of YUHL681-21 reported that her hearing loss had commenced after the administration of antibiotics for a bacterial infection, suggesting the possibility of aminoglycoside-induced hearing loss. No other syndromic symptoms were observed in any of the four patients. Segregation analysis by Sanger sequencing confirmed that the m.7444G>A variant from YUHL58 was inherited from her mother (YUHL58-12), and her brother (YUHL58-22) carried an identical mtDNA variant, suggesting maternal inheritance (Fig. 4b). Furthermore, the presence of the m.1555A>G variant in the three patients and its co-segregation in a sister (YUHL847-22) of YUHL847-21 were confirmed by segregation analysis (Fig. 4c). To further determine the genetic cause of hearing loss in these four families, we also examined other variants in hearing-loss-associated genes detected by WGS and found no pathogenic or likely pathogenic variants other than m.7444G>A or m.1555A>G variant that could plausibly explain their hearing loss. Therefore, WGS revealed deafness-causing mtDNA variants, which are typically undetectable by WES.Read-depth-based CNV analysis of WGS revealed the presence of a biallelic gross deletion and a pathogenic variantCNVs are suspected as an underrecognized cause of nonsyndromic hearing loss38. Because the sensitivity and coverage, essential for CNV discovery, of WGS have been previously reported to surpass those of WES39, we comprehensively analyzed CNVs within hearing-loss-associated loci using WGS (Supplementary Fig. 4). We identified several gross deletions in causative genes for nonsyndromic hearing loss, including STRC, COL1A1, CHD7, PTPRQ, SLC12A2 and ABHD12 (Supplementary Table 6). Among these, pathogenic CNVs in chromosomal position 15q15.3, known as the most common loci harboring long deletions or CNVs in patients with nonsyndromic hearing loss40, were identified in YUHL471-21 and YUHL488-21 (Fig. 5a).Fig. 5: WGS revealed that a gross deletion in STRC led to early-onset, moderate and recessive hearing loss.a A Circos plot of the human whole genome and a magnified region of chromosome 15 including the STRC gene. The innermost part of the circle represents genes associated with hearing loss. The scatter plot and heat map represent the read depth of detected CNVs from the analysis. Green color represents duplication, and red color represents deletion. Putative pathogenic homozygous deletion of YUHL471-21 and heterozygous deletion of YUHL488-21 in the STRC region are plotted as blue stars. The histogram represents the number of patients harboring variants present in a region when the chromosome is divided into 2,000-kbp bins. b, c A pedigree (b) and PTA result (c) of YUHL471-21. AC, air conduction thresholds; BC, bone conduction thresholds. d Visualization of read-depth (RD) signals of detected homozygous deletion in the YUHL471 family. The deleted region is bordered by two light-blue lines. The red line indicates average read-depth signals spanning a given chromosomal position. e, f A pedigree (e) and PTA result (f) of YUHL488-21. AC, air conduction thresholds; BC, bone conduction thresholds. g Visualization of RD signals of detected heterozygous deletion in the YUHL488 family. The deleted region is bordered by two light-blue lines. The red line indicates average read-depth signals spanning a given chromosomal position. Heterozygous deletion in the STRC gene was paternally inherited. h Segregation results of a c.4309C>T variant in the YUHL488 family. Sanger sequencing chromatograms revealed that heterozygous c.4309C>T was inherited from the mother of a proband and c.4309C>T was sequenced as a hemizygote in the proband due to a heterozygous deletion inherited from the father. -21, proband; -11, father; -12, mother.Full size imageYUHL471-21 was a 6-year-old female with hearing loss that began between the ages of 2 and 3 years. The family history indicated that her hearing loss was autosomal recessive, and pure-tone audiometry revealed bilateral, moderate hearing loss with a flat configuration (Fig. 5b, c). We conducted WGS for this patient to identify the molecular cause of her hearing loss, and read-depth-based CNV analysis by CNVnator revealed homozygous deletions of the STRC and CATSPER2 genes in YUHL471-21 (Fig. 5d and Supplementary Fig. 5). Nucleotide resolution of breakpoints was apparent in this complete deletion region (Supplementary Fig. 5). Moreover, heterozygous deletion of the identical region in her parents (YUHL471-11 and YUHL471-12) was also confirmed on the basis of the read-depth signal (Fig. 5d and Supplementary Fig. 5). Based on the results of trio-analysis and typical clinical characteristics of STRC-associated hearing loss, which is the most common cause of moderate, flat-configured hearing loss in Korean individuals41, we determined that the hearing loss of YUHL471-21 was attributable to the homozygous deletion of STRC.Another patient, YUHL488-21, detected with a CNV of STRC, was a 17-year-old male referred for hearing loss that started in his early teenage years. His hearing loss appeared to be autosomal recessive, and audiologic examination showed bilateral, moderate and flat-configured SNHL (Fig. 5e, f). YUHL488-21 had an extensive 26,000-bp heterozygous deletion in the CKMT1B-STRC gene, and the deletion in the same region was also detected in the father of the proband, YUHL488-11 (Fig. 5g and Supplementary Table 6). The read-depth estimated by the CNV detection tool confirmed that this deletion was heterozygous in YUHL488-21 and 488-11, compared with the mother who did not have this CNV (Fig. 5g and Supplementary Fig. 5). Interestingly, the nonsense variant, c.4309C>T, located in the aforementioned region of deletion in STRC, was detected as hemizygous in the proband (Fig. 5e, h). Subsequent segregation analysis revealed that this variant was detected as heterozygous in the mother, YUHL488-12 (Fig. 5h), and was maternally inherited. The c.4309C>T variant was reported to be rare (with an allele frequency of 0.000004959, according to gnomAD), and it creates a stop codon at a conserved glutamine residue, resulting in a truncated protein that is 339 amino acids shorter than the full-length protein and may impair its normal function. Considering the autosomal recessive inheritance pattern of hearing loss within the family, the stereotypic clinical characteristics of STRC-related hearing loss42, and the molecular characteristics of both variants, the proband’s hearing loss probably resulted from a concurrent nonsense mutation (c.4309C>T) and heterozygous copy number deletion in the STRC gene.The random insertion of TEs into a gene may contribute to the disruption of its functionTEs, particularly retrotransposons, are segments of DNA that can propagate within the genome and are largely divided into three groups: long interspersed nuclear elements 1 (L1), Alu and SINE-VNTR-Alu43. TEs have biological significance by potentially acting as insertional mutagens and perturbing the arrangement of DNA, resulting in the alteration of gene expression or RNA splicing44,45. In this regard, we thoroughly inspected the insertion of TEs in known hearing-loss-associated loci using WGS samples. After removing concurrent intersample TEs and considering allele frequencies retrieved from the population database, we finally selected several potentially detrimental insertions of TEs in SLC17A8, CDH23, EYA1, PAX3 and TMC1 (Supplementary Table 7).Among the identified patients, YUHL1109-21 was a 51-year-old female with progressive and moderate SNHL in both ears that started in her early 30s (Fig. 6a). Her mother and brothers also had hearing loss, suggesting an autosomal dominant inheritance pattern (Fig. 6b). We conducted WGS to identify the molecular cause of her hearing loss, and no pathogenic or likely pathogenic variants were identified in the coding regions of hearing-loss-associated genes. However, in-depth analysis of WGS data to identify TEIs discovered chimeric reads at the chr12:100371388 (GRCh38) position due to a heterozygous Alu insertion in the 1st intron of the SLC17A8 gene in the proband, characterized by nearby soft-clipped reads, unmapped poly-A tails and targeted-site duplications (Fig. 6c). According to the gnomAD-SV database, this insertion has not been documented in the general population. When the genomic region of the TEI was amplified by PCR, an abnormal PCR product with a larger size than the normal product was observed in YUHL1109-21 (Fig. 6d), corroborating the possibility of a heterozygous Alu insertion. Although the postlingual, progressive and high-frequency affected hearing loss of YUHL1109-21 was consistent with previous reports of SLC17A8-associated hearing loss46, whether this Alu insertion in the SLC17A8 gene functionally contributes to the disease requires further investigation.Fig. 6: TE Alu insertion was identified in SLC17A8 of patients with hearing loss by WGS screening.a, b PTA results (a) and a pedigree (b) of YUHL1109-21, AC, air conduction thresholds; BC, bone conduction thresholds. c An Integrative Genomics Viewer image at the insertion site in gene SLC17A8 in YUHL1109-21. The sequencing coverage (top) and sequencing reads (bottom) are shown. The insertion shows the canonical signatures of target-primed reverse transcription (TPRT)-mediated retrotransposition: 15-bp target site duplication (TSD) between the two insertion breakpoints, a poly-A tail, supporting clipped reads and discordant reads with mates mapping to the consensus Alu sequence. d Full-length PCR validation of the Alu insertion in SLC17A8 in YUHL1109-21. In blood-derived genomic DNA, we validated only the insertions in the proband, compared with the nonaffected control.Full size imageMutations in cis-regulatory elements might contribute to nonsyndromic hearing lossThe precise spatiotemporal regulation of gene expression plays a critical role in the normal development of the inner ear from the otocyst47. Therefore, we hypothesized that disruptions in the timely orchestrated regulation of gene expression, stemming from overlooked variants in cis-regulatory elements of genes associated with hearing loss, could potentially contribute to the condition. Utilizing data from phase III of the ENCODE Project22, we screened for variants that could potentially affect the expression of known hearing-loss-associated genes, among 926,535 human cCREs from the SCREEN database22. Among these elements, we identified six potential variants in cis-regulatory elements associated with hearing loss genes, CDH23, OTOF, MYO6, COL4A4 and WHRN (Supplementary Table 8). Considering the inheritance pattern and clinical phenotypes, we determined that these variants are insufficient to be classified as pathogenic. Nevertheless, our results emphasize the need to identify potential mutations in regulatory regions of known hearing loss genes and validate their effects on the elaborate regulation of inner-ear-specific transcription factors to reveal unknown causes of hearing loss.Secondary findings in patients with deafness by WGSIn addition to genetic factors, congenital CMV infection is also one of the leading causes of SNHL and neurodevelopmental complications48. Generally, newborns are screened for CMV infection by PCR of CMV DNA in urine or saliva, but a recent study detected CMV infection by utilizing WGS from blood specimens49. Therefore, we screened for potential CMV infection in our patients with hearing loss who underwent WGS. Consequently, we found that a significant number of reads were aligned to the CMV genome among 1.4% (2/140) of patients, suggesting a history of CMV infection (Supplementary Table 9). However, both identified patients already carried causative pathogenic variants in LOXHD1 and KCNQ4 attributable to their hearing loss (Supplementary Table 9). Because hearing loss arises from congenital CMV infection, in contrast to postnatal CMV infection, which is not recognized as a direct cause of SNHL in healthy individuals50, and WGS cannot determine the exact period of infection, we could not ascertain that the identified evidence of CMV infection from WGS is the cause of their hearing loss. However, these cases suggest that WGS can provide additional valuable information on one of the overlooked causes of SNHL, alongside genetic etiologies.Recently, the American College of Medical Genetics and Genomics published guidance for reporting secondary findings in the context of clinical exome and genome sequencing51. For opportunistic screening to identify and manage additional risks for specific genetic disorders not just limited to SNHL and offer established interventions to patients, we detected variants of these listed genes classified as pathogenic or likely pathogenic in the clinical database. This approach detected several pathogenic variants in MUTYH, MSH2, TTN, DES, BTD, BRCA2, ATP7B, TP53, GAA and LDLR genes (Supplementary Table 10). Most of variants were loss-of-function variants perturbing protein, such as nonsense, frameshift and aberrant splice variants, located in obligatory splicing sites. Interestingly, the c.744+2del variant in MSH2 was highly frequent in our cohort. Further clinical interpretation of these secondary findings is required for critical discussions with clinicians considering ethnicity-specific frequencies of allele, inheritance mode and phenotype of patients.DiscussionIn recent years, next-generation sequencing has undergone substantial advances, proving its diagnostic value in unveiling the genetic etiologies of multiple disease52. In the realm of hereditary hearing loss, multiple studies utilizing high-throughput WES have identified key causative genes contributing significantly to hearing loss11,12,13,14,15,16. However, despite the advantages of WES, over 60% of patients with Mendelian inheritance patterns of hearing loss remain undiagnosed, lacking precise molecular characterization (Fig. 1). This lack of conclusive genetic diagnoses arises from various factors, including ambiguous genotype–phenotype correlations, insufficient evidence supporting identified variants, incomplete penetrance, variable expressivity or allelic heterogeneity in hearing loss17. Moreover, from a technical perspective, WES may fail to detect variants located outside coding regions, such as those in intergenic or deep-intronic regions, cis-regulatory regions or mtDNA, and it may not be suitable for sensitively identifying CNVs. Indeed, recent study in families with rare monogenic diseases suggested that approximately 8% (61/744) of diagnoses were missed by exome sequencing, but when the genetic screening method was expanded to WGS, the diagnostic yield of those negative evaluations increased by 2.3% (17/744)53. This result highlights the need to utilize platforms such as short-read genome sequencing or even long-read genome sequencing for the diagnosis of variants that may elude exome sequencing. Accordingly, to address limitations of WES or targeted sequencing and elucidate the genetic etiologies of unresolved cases by WES, we conducted extensive WGS analyses on patients with nonsyndromic hearing loss and established a systematic pipeline for investigating pathogenic genetic alterations throughout the entire genome.The discovery of putative pathogenic intronic variants that disrupt normal splicing was one of the important findings from our study. During variant discovery using WES, variants located outside of obligatory splicing sites are frequently overlooked. However, those variants located in deep-intronic regions missed by WES sometimes significantly contribute to Mendelian disorders54. Furthermore, splicing-affecting variants located in deep-intronic regions are more amenable targets of splice-switching oligonucleotide therapy than those in canonical splicing sites55. Therefore, it is imperative to identify hidden splicing-disruptive variants commonly underdetected by WES with WGS in terms of actionability for the development of novel therapeutics for hearing loss.Although less prevalent, there is a growing body of studies reporting hearing loss caused by mtDNA variants, which account for approximately 1.5–5% of hearing loss cases, depending on ethnicity37. Notably, the clinical characteristics of hearing loss caused by mtDNA variants are highly heterogeneous, and the family history of affected patients is often inconclusive due to incomplete penetrance37. Therefore, a systemic approach using WGS holds implications for discovering unsuspected and undetected mtDNA variants, particularly when the diagnosis relies solely on clinical findings and WES. Furthermore, because the m.1555A>G variant, the most prevalent mtDNA variant associated with hearing loss, is linked to augmented aminoglycoside-related ototoxicity35, accurate identification of hearing-loss-related mtDNA variants carries clinical significance and can guide physicians in preventing further disease exacerbation.CNVs have been reported as a significant cause of multiple Mendelian diseases56, including nonsyndromic hearing loss38. Especially for hearing loss, STRC is the most frequently identified causative gene with a segmental deletion of DFNB16 locus, sometimes accompanied by a deletion of adjacent CATSPER2 gene resulting in deafness–infertility syndrome57. Here, we identified homozygous deletion and heterozygous deletion with a hemizygous nonsense mutation in the STRC gene and successfully mapped the exact breakpoint of deletion with WGS data. The mapping of breakpoints of CNVs can enable the interpretation of structural complexity and pathogenicity of identified CNVs, which is challenging with WES data providing only low-resolution information due to limited genome coverage and depth58. In particular, the aforementioned hemizygous nonsense variant was initially not detected in WES but was identified upon reexamining the data with WGS. Furthermore, besides deletions in the STRC gene, CNVs in other deafness-causative genes, such as OTOA and TMC1, have been reported38, underscoring the importance of identifying CNVs and interpreting their impact on the pathogenesis of the disease.The inner ear is a complex organ comprising highly heterogeneous cell types, each differing in morphology, functional role and anatomical localization. The tightly orchestrated differentiation and development of these various cell types are governed by multiple key transcription factors dictating gene regulatory networks in a timely manner47,59. However, while the identification of variants in cis-regulatory regions has been reported in multiple other diseases60, their exact contribution to hearing loss remains largely unknown. Therefore, the detection of variants in cis-regulatory regions of multiple deafness-causative genes in our study underscores the need for multi-omics analysis to elucidate the pathogenic role of these variants in hearing loss. Although our study did not identify any pathogenic cCRE variants, we assert that a more comprehensive analysis of the association between these cCRE variants and hearing loss phenotypes, utilizing approaches such as burden test or association-based analysis, is necessary, contingent upon the availability of proper control.Recently, TEs, particularly retrotransposons, have been increasingly recognized as significant contributors to Mendelian diseases. Furthermore, through WGS, detrimental insertions of TE into deep-intronic regions have been reported, and their impacts on normal splicing and therapeutic implications have also been unveiled44,61. In this study, we identified a hidden Alu insertion in the SLC17A8 gene by WGS in a patient with hearing loss otherwise unexplained by WES. A more accurate estimation of the contribution of mobile element insertion to hearing loss requires further analysis with a larger cohort.In addition to the identification of potential candidate variants in hearing-loss-associated genes, we also found a few deleterious variants in a list of genes with secondary findings in clinical exome and genome sequencing provided by the American College of Medical Genetics and Genomics51. These secondary findings might provide additional information to the patients of medically actionable variants and can be useful within the context of optimizing potential medical benefit to the patient, even when they are unrelated to the primary medical reason for testing.Nevertheless, this study has several limitations. Although WGS partially addressed undiagnosed cases missed by WES, the overall diagnostic rate of WGS did not exhibit a significant increase compared with that of WES. This outcome may stem from the differences in the patient populations undergoing WES versus WGS. Variations in the onset, severity and mode of inheritance of hearing loss were observed between patients subjected to WES and those undergoing WGS. Patients undergoing WES displayed a higher frequency of postlingual, autosomal dominant and multiplex hearing loss cases, potentially impacting the diagnostic yield. In addition, this study did not functionally confirm the contribution of TEIs to the pathology. Further investigations, such as minigene assays to evaluate the impact of mobile element insertions on normal splicing, and transcriptomic analyses62 to assess the consequences of variants on gene regulatory networks of specific cell types within the inner ear, are necessary to determine whether these variants indeed induce hearing loss.Overall, this study presented previously overlooked genetic etiologies of hearing loss through a comprehensive systemic analysis based on WGS. In addition to coding region variants, intronic variants disrupting normal splicing, mtDNA variants and CNVs were identified in patients with nonsyndromic hearing loss unresolved by WES. Furthermore, TEIs potentially contributing to hearing loss were uncovered. Through a thorough investigation into the molecular basis of hearing loss with WGS, a complete genetic landscape of the condition will be unveiled. Furthermore, based on these findings, the development of patient-customized therapeutics will become feasible.
Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.
ReferencesKorver, A. M. et al. Congenital hearing loss. Nat. Rev. Dis. Primers 3, 16094 (2017).PubMed
PubMed Central
Google Scholar
Cunningham, L. L. & Tucci, D. L. Hearing loss in adults. N. Engl. J. Med. 377, 2465–2473 (2017).PubMed
PubMed Central
Google Scholar
Han, J. H. et al. Comprehensive prediction model, including genetic testing, for the outcomes of cochlear implantation. Ear Hear. 44, 223–231 (2023).PubMed
Google Scholar
Azaiez, H. et al. Genomic landscape and mutational signatures of deafness-associated genes. Am. J. Hum. Genet. 103, 484–497 (2018).CAS
PubMed
PubMed Central
Google Scholar
Guan, Q. et al. AUDIOME: a tiered exome sequencing-based comprehensive gene panel for the diagnosis of heterogeneous nonsyndromic sensorineural hearing loss. Genet. Med. 20, 1600–1608 (2018).PubMed
Google Scholar
Shearer, A. E. et al. Advancing genetic testing for deafness with genomic technology. J. Med. Genet. 50, 627–634 (2013).CAS
PubMed
Google Scholar
Downie, L. et al. Exome sequencing in infants with congenital hearing impairment: a population-based cohort study. Eur. J. Hum. Genet. 28, 587–596 (2020).CAS
PubMed
Google Scholar
Wu, J. et al. Molecular diagnose of a large hearing loss population from China by targeted genome sequencing. J. Hum. Genet. 67, 643–649(2022).CAS
PubMed
PubMed Central
Google Scholar
Molina-Ramírez, L. P. et al. The diagnostic utility of clinical exome sequencing in 60 patients with hearing loss disorders: a single-institution experience. Clin. Otolaryngol. 46, 1257–1262 (2021).PubMed
Google Scholar
Baux, D. et al. Combined genetic approaches yield a 48% diagnostic rate in a large cohort of French hearing-impaired patients. Sci. Rep. 7, 16783 (2017).CAS
PubMed
PubMed Central
Google Scholar
Oh, K. S. et al. Overlooked KCNQ4 variants augment the risk of hearing loss. Exp. Mol. Med. 55, 844–859 (2023).CAS
PubMed
PubMed Central
Google Scholar
Joo, S. Y. et al. Clinical heterogeneity associated with MYO7A variants relies on affected domains. Biomedicines https://doi.org/10.3390/biomedicines10040798 (2022).Na, G. et al. Heterogeneity of MYO15A variants significantly determine the feasibility of acoustic stimulation with hearing aid and cochlear implant. Hear. Res. 404, 108227 (2021).PubMed
Google Scholar
Koh, Y. I. et al. OSBPL2 mutations impair autophagy and lead to hearing loss, potentially remedied by rapamycin. Autophagy 18, 2593–2614 (2022).CAS
PubMed
PubMed Central
Google Scholar
Kang, M. et al. Novel variant in CEP250 causes protein mislocalization and leads to nonsyndromic autosomal recessive type of progressive hearing loss. Cells https://doi.org/10.3390/cells12182328 (2023).Oh, K. S. et al. COCH-related autosomal dominant nonsyndromic hearing loss: a phenotype–genotype study. Hum. Genet. 141, 889–901 (2022).PubMed
Google Scholar
Wojcik, M. H. et al. Beyond the exome: what’s next in diagnostic testing for Mendelian conditions. Am. J. Hum. Genet. 110, 1229–1248 (2023).CAS
PubMed
PubMed Central
Google Scholar
Palmer, E. E. et al. Diagnostic yield of whole genome sequencing after nondiagnostic exome sequencing or gene panel in developmental and epileptic encephalopathies. Neurology 96, e1770–e1782 (2021).CAS
PubMed
Google Scholar
Schon, K. R. et al. Use of whole genome sequencing to determine genetic basis of suspected mitochondrial disorders: cohort study. BMJ 375, e066288 (2021).PubMed
PubMed Central
Google Scholar
Ramzan, M. et al. Genome sequencing identifies coding and non-coding variants for non-syndromic hearing loss. J. Hum. Genet. https://doi.org/10.1038/s10038-023-01159-9 (2023).Belkadi, A. et al. Whole-genome sequencing is more powerful than whole-exome sequencing for detecting exome variants. Proc. Natl Acad. Sci. USA 112, 5473–5478 (2015).CAS
PubMed
PubMed Central
Google Scholar
Moore, J. E. et al. Expanded encyclopaedias of DNA elements in the human and mouse genomes. Nature 583, 699–710 (2020).PubMed
PubMed Central
Google Scholar
Park, H. J., Hahn, S. H., Chun, Y. M., Park, K. & Kim, H. N. Connexin26 mutations associated with nonsyndromic hearing loss. Laryngoscope 110, 1535–1538 (2000).CAS
PubMed
Google Scholar
Park, H. J. et al. Origins and frequencies of SLC26A4 (PDS) mutations in east and south Asians: global implications for the epidemiology of deafness. J. Med. Genet. 40, 242–248 (2003).CAS
PubMed
PubMed Central
Google Scholar
Fadaie, Z. et al. Whole genome sequencing and in vitro splice assays reveal genetic causes for inherited retinal diseases. NPJ Genom. Med. 6, 97 (2021).CAS
PubMed
PubMed Central
Google Scholar
Jaganathan, K. et al. Predicting splicing from primary sequence with deep learning. Cell 176, 535–548 (2019).CAS
PubMed
Google Scholar
Unzaki, A. et al. Clinically diverse phenotypes and genotypes of patients with branchio-oto-renal syndrome. J. Hum. Genet. 63, 647–656 (2018).CAS
PubMed
Google Scholar
Chang, E. H. et al. Branchio-oto-renal syndrome: the mutation spectrum in EYA1 and its phenotypic consequences. Hum. Mutat. 23, 582–589 (2004).CAS
PubMed
Google Scholar
Nam, D. W. et al. Molecular genetic etiology and revisiting the middle ear surgery outcomes of branchio-oto-renal syndrome: experience in a tertiary referral center. Otol. Neurotol. 44, e319–e327 (2023).PubMed
Google Scholar
Vaché, C. et al. Nasal epithelial cells are a reliable source to study splicing variants in Usher syndrome. Hum. Mutat. 31, 734–741 (2010).PubMed
Google Scholar
Valero, R. et al. Aberrant splicing events associated to CDH23 noncanonical splice site mutations in a proband with atypical Usher syndrome 1. Genes https://doi.org/10.3390/genes10100732 (2019).Wagatsuma, M. et al. Distribution and frequencies of CDH23 mutations in Japanese patients with non-syndromic hearing loss. Clin. Genet. 72, 339–344 (2007).CAS
PubMed
Google Scholar
Yuan, H. et al. Cosegregation of the G7444A mutation in the mitochondrial COI/tRNA(Ser(UCN)) genes with the 12S rRNA A1555G mutation in a Chinese family with aminoglycoside-induced and nonsyndromic hearing loss. Am. J. Med. Genet. A 138a, 133–140 (2005).PubMed
PubMed Central
Google Scholar
Pandya, A. et al. Heterogenous point mutations in the mitochondrial tRNA Ser(UCN) precursor coexisting with the A1555G mutation in deaf students from Mongolia. Am. J. Hum. Genet. 65, 1803–1806 (1999).CAS
PubMed
PubMed Central
Google Scholar
Prezant, T. R. et al. Mitochondrial ribosomal RNA mutation associated with both antibiotic-induced and non-syndromic deafness. Nat. Genet. 4, 289–294 (1993).CAS
PubMed
Google Scholar
Bae, J. W. et al. Molecular and clinical characterization of the variable phenotype in Korean families with hearing loss associated with the mitochondrial A1555G mutation. PLoS ONE 7, e42463 (2012).CAS
PubMed
PubMed Central
Google Scholar
Joo, S. Y. et al. Prevalence and clinical characteristics of mitochondrial DNA mutations in Korean patients with sensorineural hearing loss. J. Korean Med. Sci. 38, e355 (2023).CAS
PubMed
PubMed Central
Google Scholar
Shearer, A. E. et al. Copy number variants are a common cause of non-syndromic hearing loss. Genome Med. 6, 37 (2014).PubMed
PubMed Central
Google Scholar
Sun, Y. et al. Characterizing sensitivity and coverage of clinical WGS as a diagnostic test for genetic disorders. BMC Med. Genomics 14, 102 (2021).CAS
PubMed
PubMed Central
Google Scholar
Nishio, S. Y. & Usami, S. I. Frequency of the STRC-CATSPER2 deletion in STRC-associated hearing loss patients. Sci. Rep. 12, 634 (2022).CAS
PubMed
PubMed Central
Google Scholar
Kim, B. J. et al. Significant Mendelian genetic contribution to pediatric mild-to-moderate hearing loss and its comprehensive diagnostic approach. Genet. Med. 22, 1119–1128 (2020).CAS
PubMed
Google Scholar
Back, D. et al. Phenotypic characterization of DFNB16-associated hearing loss. Otol Neurotol 40, e48–e55 (2019).PubMed
Google Scholar
Garret, P. et al. A second look at exome sequencing data: detecting mobile elements insertion in a rare disease cohort. Eur. J. Hum. Genet. https://doi.org/10.1038/s41431-022-01250-3 (2022).Kim, J. et al. Patient-customized oligonucleotide therapy for a rare genetic disease. N. Engl. J. Med. 381, 1644–1652 (2019).CAS
PubMed
PubMed Central
Google Scholar
Hancks, D. C. & Kazazian, H. H. Jr Roles for retrotransposon insertions in human disease. Mob. DNA 7, 9 (2016).PubMed
PubMed Central
Google Scholar
Ryu, N. et al. Screening of the SLC17A8 gene as a causative factor for autosomal dominant non-syndromic hearing loss in Koreans. BMC Med. Genet. 17, 6 (2016).PubMed
PubMed Central
Google Scholar
Jan, T. A. et al. Spatiotemporal dynamics of inner ear sensory and non-sensory cells revealed by single-cell transcriptomics. Cell Rep. 36, 109358 (2021).CAS
PubMed
PubMed Central
Google Scholar
Singh, G. & Gaidhane, A. A review of sensorineural hearing loss in congenital cytomegalovirus infection. Cureus 14, e30703 (2022).PubMed
PubMed Central
Google Scholar
Xiang, J. et al. Utility of whole genome sequencing for population screening of deafness-related genetic variants and cytomegalovirus infection in newborns. Front. Genet. 13, 883617 (2022).PubMed
PubMed Central
Google Scholar
Aldè, M. et al. Congenital cytomegalovirus and hearing loss: the state of the art. J. Clin. Med. https://doi.org/10.3390/jcm12134465 (2023).Miller, D. T. et al. ACMG SF v3.2 list for reporting of secondary findings in clinical exome and genome sequencing: a policy statement of the American College of Medical Genetics and Genomics (ACMG). Genet. Med. 25, 100866 (2023).CAS
PubMed
PubMed Central
Google Scholar
Yang, Y. et al. Clinical whole-exome sequencing for the diagnosis of mendelian disorders. N. Engl. J. Med. 369, 1502–1511 (2013).CAS
PubMed
PubMed Central
Google Scholar
Wojcik, M. H. et al. Genome sequencing for diagnosing rare diseases. N. Engl. J. Med. 390, 1985–1997 (2024).CAS
PubMed
PubMed Central
Google Scholar
Reurink, J. et al. Whole genome sequencing for USH2A-associated disease reveals several pathogenic deep-intronic variants that are amenable to splice correction. HGG Adv. 4, 100181 (2023).CAS
PubMed
PubMed Central
Google Scholar
Kim, J. et al. A framework for individualized splice-switching oligonucleotide therapy. Nature 619, 828–836 (2023).CAS
PubMed
PubMed Central
Google Scholar
Stankiewicz, P. & Lupski, J. R. Structural variation in the human genome and its role in disease. Annu. Rev. Med. 61, 437–455 (2010).CAS
PubMed
Google Scholar
Zhang, Y. et al. Sensorineural deafness and male infertility: a contiguous gene deletion syndrome. BMJ Case Rep. https://doi.org/10.1136/bcr.08.2008.0645 (2009).Abbasi, W., French, C. E., Rockowitz, S., Kenna, M. A. & Eliot Shearer, A. Evaluation of copy number variants for genetic hearing loss: a review of current approaches and recent findings. Hum. Genet. 141, 387–400 (2022).PubMed
Google Scholar
Shin, J. O. et al. CTCF regulates otic neurogenesis via histone modification in the Neurog1 locus. Mol. Cells 41, 695–702 (2018).CAS
PubMed
PubMed Central
Google Scholar
Thomas, E. D. et al. Cell-specific cis-regulatory elements and mechanisms of non-coding genetic disease in human retina and retinal organoids. Dev. Cell 57, 820–836 (2022).CAS
PubMed
PubMed Central
Google Scholar
Zhao, B. et al. Contribution and therapeutic implications of retroelement insertions in ataxia telangiectasia. Am. J. Hum. Genet. 110, 1976–1982 (2023).CAS
PubMed
PubMed Central
Google Scholar
Lee, G. Y., Ham, S. & Lee, S. V. Brief guide to RNA sequencing analysis for nonexperts in bioinformatics. Mol. Cells 47, 100060 (2024).CAS
PubMed
PubMed Central
Google Scholar
Download referencesAcknowledgementsWe extend our sincere appreciation to the families for their agreement to participate in this study. We also express our gratitude to the Korean population database for this study provided by the National Biobank of Korea, Centers for Disease Control and Prevention, Republic of Korea (4845-301, 4851-302 and 4851-307). This work was supported by the National Research Foundation of Korea (NRF) grants funded by the Korean government (RS-2024-00438709 and RS-2023-00261905 to H.Y.G. and RS-2024-00346485 to J.J.), the Team Science Award of Yonsei University College of Medicine (6-2021-0002 to J.J. and 6-2021-0003 to H.Y.G.) and the MD-PhD/Medical Scientist Training Program through the Korea Health Industry Development Institute (KHIDI) to S.H.J., funded by the Ministry of Health and Welfare, Republic of Korea.Author informationAuthor notesThese authors contributed equally: Jung Ah Kim, Seung Hyun Jang.Authors and AffiliationsDepartment of Pharmacology, Graduate School of Medical Science, Brain Korea 21 Project, Yonsei University College of Medicine, Seoul, Republic of KoreaJung Ah Kim, Seung Hyun Jang, Sun Yung Joo, Se Jin Kim & Heon Yung GeeWon-Sang Lee Institute for Hearing Loss, Seoul, Republic of KoreaJung Ah Kim, Seung Hyun Jang, Sun Yung Joo, Se Jin Kim, Jae Young Choi, Jinsei Jung & Heon Yung GeeDepartment of Otorhinolaryngology, Yonsei University College of Medicine, Seoul, Republic of KoreaJae Young Choi & Jinsei JungAuthorsJung Ah KimView author publicationsYou can also search for this author inPubMed Google ScholarSeung Hyun JangView author publicationsYou can also search for this author inPubMed Google ScholarSun Yung JooView author publicationsYou can also search for this author inPubMed Google ScholarSe Jin KimView author publicationsYou can also search for this author inPubMed Google ScholarJae Young ChoiView author publicationsYou can also search for this author inPubMed Google ScholarJinsei JungView author publicationsYou can also search for this author inPubMed Google ScholarHeon Yung GeeView author publicationsYou can also search for this author inPubMed Google ScholarCorresponding authorsCorrespondence to
Jinsei Jung or Heon Yung Gee.Ethics declarations
Competing interests
The authors declare no competing interests.
Additional informationPublisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Supplementary informationSupplementary InformationRights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Reprints and permissionsAbout this articleCite this articleKim, J.A., Jang, S.H., Joo, S.Y. et al. Systematic genetic assessment of hearing loss using whole-genome sequencing identifies pathogenic variants.
Exp Mol Med (2025). https://doi.org/10.1038/s12276-025-01428-xDownload citationReceived: 01 August 2024Revised: 20 December 2024Accepted: 13 January 2025Published: 01 April 2025DOI: https://doi.org/10.1038/s12276-025-01428-xShare this articleAnyone you share the following link with will be able to read this content:Get shareable linkSorry, a shareable link is not currently available for this article.Copy to clipboard
Provided by the Springer Nature SharedIt content-sharing initiative