Mutation Spectrum and Novel Rare Variants in a Han Chinese Hearing Loss Cohort Using a Tiered Sequencing Strategy Authors: Yu Zhang, Shimei Yuan, Yafang Wan, Tian Li, Lanlan Xu, Wei Yang, Changchun Niu Published: 2026 ABSTRACT ABSTRACTBackgroundHereditary hearing loss is a genetically heterogeneous disorder posing challenges for molecular diagnosis. This study aimed to characterize the mutation spectrum, including rare and previously unreported variants, using a tiered strategy in a Han Chinese cohort from Southwest China.MethodsIn this retrospective study, 43 patients underwent targeted hotspot screening, followed by whole‐exome sequencing in 11 patients with no hotspot variants detected. Variants were annotated and interpreted according to American College of Medical Genetics and Genomics guidelines.ResultsInitial hotspot screening established molecular diagnoses in 23 patients based on seven pathogenic or likely pathogenic variants in MT‐RNR1, GJB2, and SLC26A4. Whole‐exome sequencing identified six additional rare variants in CDH23, ILDR1, OTOF, and SLC12A2, and established two diagnoses involving ILDR1 and CDH23. A previously unreported SLC12A2 missense variant, c.2936A>G (p.Glu979Gly) was identified as a variant of uncertain significance and retained as a candidate finding. CDH23 c.5311C>T (p.Arg1771Ter) has not previously been reported in Han Chinese patients with hearing loss, to our knowledge.ConclusionsThese findings expand the mutation spectrum of hereditary hearing loss in a Han Chinese cohort from Southwest China and highlight the practical value of a tiered testing strategy integrating targeted hotspot screening with whole‐exome sequencing.Keywords: Han Chinese, mutation spectrum, rare variants, tiered sequencing strategy, whole‐exome sequencing Background Hearing loss (HL) is a common condition with major global impact, affecting more than 1.5 billion people worldwide (World Health Organization 2021). Genetic factors account for over 60% of cases (Yuan et al. 2020), and more than 220 genes associated with deafness have been identified (MORL, n.d.). The prevalence and spectrum of pathogenic variants differ across ethnic groups, with particularly high carrier rates reported in the Chinese population (Feng et al. 2024; Hou et al. 2024), and HL also shows marked genetic heterogeneity (Koffler et al. 2015). In routine clinical practice, PCR‐hybridization assays remain the main screening tool in many regions of China, including the Southwest, because of their low cost and ease of use. However, these assays typically target only a limited set of mutation hotspots. Consequently, they often fail to capture rare variants, leaving a proportion of patients without a definitive molecular explanation. The advent of next‐generation sequencing (NGS), including targeted panels and whole‐exome sequencing (WES), has facilitated the identification of rare genetic etiologies (Wu et al. 2022; Zeng et al. 2022). While tiered strategies—utilizing hotspot screening followed by WES—have been proposed to balance efficiency and depth, population‐specific mutational data for the Han Chinese population in Southwest China remain sparse. In this study, we applied a tiered sequencing approach to investigate the molecular landscape of HL in a Han Chinese cohort. By combining hotspot panel testing with WES, we aimed to characterize the mutation spectrum and identify novel or rare genetic contributors, thereby providing population‐specific data for this region. Methods Subjects A cohort of 43 patients with hearing loss was recruited from the central region of Chongqing, Southwest China. Patients with clinical features suggestive of syndromic hearing loss were not included. Clinical information, including age at onset, family history, and history of ototoxic drug exposure, was collected through patient interviews and review of available medical records. Patients were classified according to the reported age at onset of HL. In addition, available family members were recruited for genetic analysis. Peripheral blood (3–5 mL) was collected in EDTA tubes, and genomic DNA was extracted using standard protocols. Genetic Sequencing and Analysis Initially, all 43 patients underwent a high‐throughput targeted screening assay covering 208 common deafness‐associated variants across 24 genes (Table S1), designed to detect prevalent hotspots in the Chinese population. Subsequently, WES was performed on a selected subset of 11 patients who tested negative in the initial screening. Sequencing was performed by a certified commercial service provider (BGI Genomics, Shenzhen, China) with standard protocols. Raw data were filtered and aligned to the human reference genome (hg19/GRCh37). Then, variants were annotated using ANNOVAR (Wang et al. 2010) and cross‐referenced with multiple databases, including dbSNP (build 151) (Sherry et al. 2001), gnomAD (v2.1.1) (Karczewski et al. 2020), ExAC (release 0.3.1) (Lek et al. 2016), ClinVar (accessed May 2026) (Landrum et al. 2018), and HGMD professional (accessed May 2026) (Stenson et al. 2017). Candidate protein‐altering variants were prioritized in known hearing‐loss‐associated genes and genes relevant to the HL phenotype. Prioritization also considered case‐specific clinical information, the presumed mode of inheritance, computational predictions, and published evidence. Population allele frequency was evaluated using the East Asian population dataset of gnomAD (v2.1.1). Pathogenicity assessment incorporated multiple in silico prediction tools, including MutationTaster (Steinhaus et al. 2021), REVEL (Ioannidis et al. 2016), and M‐CAP (Jagadeesh et al. 2016) for missense effect prediction. Evolutionary conservation was evaluated using GERP++ (Davydov et al. 2010) scores and multiple sequence alignment of mammalian orthologs generated with UGENE (Okonechnikov et al. 2012) software. Variant classification was determined according to the American College of Medical Genetics and Genomics and Association for Molecular Pathology (ACMG/AMP) framework (Richards et al. 2015). Results Overview of Cohort Characteristics and Initial Hotspot Screening The cohort included 43 patients with HL, comprising 30 males (69.8%), and 13 females (30.2%), with an age range of 21–53 years (mean ± SD: 39 ± 7.2 years; median: 39 years). These patients represented 37 families. Five families included two unrelated spouses, and one family included two biological brothers (P25 and P26). Except for P25 and P26, all participants were biologically unrelated. Patients were classified according to the reported age at onset of HL, including congenital, prelingual (1–3 years), childhood/adolescent (4–17 years), and adult (≥ 18 years) onset (Table 1). Sixteen patients (37.2%) reported exposure to ototoxic medication, and 14 (32.6%) had a family history of hearing loss. Detailed demographic and clinical characteristics are summarized in Table 1. TABLE 1Demographic and clinical characteristics of the cohort.Characteristic n (%)Total patients43 (100.0)SexMale30 (69.8)Female13 (30.2)Age at onset of hearing lossCongenital20 (46.5)Prelingual (1–3 years)17 (39.5)Childhood/Adolescent (4–17 year)6 (14.0)Adult (≥ 18 year)0 (0.0)Family historyPositive14 (32.6)Negative or unknown29 (67.4)Presumed mode of inheritance a Autosomal dominant1 (2.3)Autosomal recessive10 (23.3)Maternal15 (34.9)Sporadic13 (30.2)Indeterminate4 (9.3)Ototoxic drug exposureYes16 (37.2)No or unknown27 (62.8) a Presumed mode of inheritance was assigned based on the available family‐history information and, where informative, molecular findings. Cases with no other affected biological relatives were classified as sporadic. Cases with insufficient or inconclusive evidence to support a specific inheritance pattern were classified as indeterminate. Demographic and clinical characteristics of the cohort. Presumed mode of inheritance was assigned based on the available family‐history information and, where informative, molecular findings. Cases with no other affected biological relatives were classified as sporadic. Cases with insufficient or inconclusive evidence to support a specific inheritance pattern were classified as indeterminate. The overall workflow of the tiered sequencing strategy is illustrated in Figure 1. The initial hotspot screening detected 7 pathogenic or likely pathogenic (P/LP) variants in 32 of 43 patients (74.4%), of these, molecular diagnoses were established in 23 patients. These variants were localized within three well‐established genes: GJB2, SLC26A4, and MT‐RNR1. Single heterozygous variants in autosomal recessive genes (e.g., GJB2) were insufficient to establish a molecular diagnosis and are summarized in the Table S2. Patients with definitive molecular diagnoses established by the initial hotspot screening are presented in Table 2. Study flowchart. A total of 43 Han Chinese patients with hereditary hearing loss underwent targeted hotspot screening, followed by whole‐exome sequencing (WES) in cases with negative panel results. Targeted screening identified pathogenic or likely pathogenic (P/LP) variants in 32 patients, while WES of 11 panel‐negative cases identified additional rare variants, including a novel SLC12A2 variant and a CDH23 variant not previously reported in Han Chinese patients with hearing loss. Abbreviations: P/LP, pathogenic or likely pathogenic; VUS, variant of uncertain significance. TABLE 2Patients with molecular diagnoses identified by targeted hotspot panel screening.CaseSexAge at onsetPresumed mode of inheritanceGeneNucleotide changeProtein changeZygosityPathogenicityP1M4 yMaternal MT‐RNR1 m.1555A>G—HomoplasmicPathogenicP4MCongenitalMaternal MT‐RNR1 m.1555A>G—HomoplasmicPathogenicP6FCongenitalMaternal MT‐RNR1 m.1555A>G—HomoplasmicPathogenicP8MCongenitalMaternal MT‐RNR1 m.1555A>G—HomoplasmicPathogenicP12M4 yMaternal MT‐RNR1 m.1555A>G—HomoplasmicPathogenicP14M2–3 yMaternal MT‐RNR1 m.1555A>G—HomoplasmicPathogenicP15MCongenitalMaternal MT‐RNR1 m.1555A>G—HomoplasmicPathogenicP17M3 yMaternal MT‐RNR1 m.1555A>G—HomoplasmicPathogenicP24M3 yMaternal MT‐RNR1 m.1555A>G—HomoplasmicPathogenicP38M1 yMaternal MT‐RNR1 m.1555A>G—HomoplasmicPathogenicP39M2.5 yMaternal MT‐RNR1 m.1555A>G—HomoplasmicPathogenicP42M1 yMaternal MT‐RNR1 m.1555A>G—HomoplasmicPathogenicP49FCongenitalMaternal MT‐RNR1 m.1555A>G—HomoplasmicPathogenicP33M4 yMaternal MT‐RNR1 m.1555A>G—Heteroplasmic a PathogenicP2F2 yMaternal MT‐RNR1 m.1494C>T—HomoplasmicPathogenicP40M2 yMaternal MT‐RNR1 m.1555A>G—HomoplasmicPathogenic GJB2 c.109G>Ap.Val37IleHeterozygousPathogenicP9MCongenitalAR GJB2 c.235delCp.Leu79fsHomozygousPathogenicP21MCongenitalAR GJB2 c.235delCp.Leu79fsHomozygousPathogenicP23F1 yAR GJB2 c.235delCp.Leu79fsHomozygousPathogenicP10MCongenitalAR GJB2 c.109G>Ap.Val37IleHomozygousPathogenicP48F2 yAR GJB2 c.109G>Ap.Val37IleHomozygousPathogenicP20MCongenitalAR SLC26A4 c.919‐2A>G—HomozygousPathogenicP45M3 yAR SLC26A4 c.[1229C>T]; [387delC]p.[Thr410Met];[Phe130fs]Compound heterozygousPathogenicAbbreviations: AR, autosomal recessive; F, female; M, male; y, years. a Heteroplasmy level: 88.6%. Patients with molecular diagnoses identified by targeted hotspot panel screening. Abbreviations: AR, autosomal recessive; F, female; M, male; y, years. Heteroplasmy level: 88.6%. Additional Variants Identified by Whole‐Exome Sequencing Whole‐exome sequencing (WES) was performed in 11 patients in whom no variants were detected by the initial hotspot screening, according to the predefined tiered sequencing strategy. WES identified six rare variants in four individuals across four genes. Findings that established a molecular diagnosis or were retained as a candidate finding are summarized in Table 3. The unresolved WES finding that did not establish a molecular diagnosis is provided in Table S2. Patient P37 had hearing loss with onset at 5–6 years of age and carried two heterozygous ILDR1 variants, c.772C>T (p.Gln258Ter) and c.206C>A (p.Pro69His). A previously unreported variant in SLC12A2 was identified and classified as a variant of uncertain significance (VUS). In addition, a CDH23 nonsense variant was identified, which has not been reported in Han Chinese patients with hearing loss to our knowledge. TABLE 3Additional findings identified by whole‐exome sequencing in panel‐negative cases.CaseSexAge at onsetPresumed mode of inheritanceGeneTranscripts a Nucleotide changeProtein changeZygosityAF b PathogenicityDiagnostic significanceP47FCongenitalAR CDH23 NM_022124.5c.[5311C>T]; [4562A>G]p.[Arg1771*]; [Asn1521Ser] compound heterozygous 0; 0.00005119P; LPDiagnosis establishedP7FCongenitalAD SLC12A2 NM_001046.2c.2936A>Gp.Glu979Glyheterozygous—VUSCandidate findingP37M5–6 yAR ILDR1 NM_001199799.1c.[772C>T]; [206C>A]p.[Gln258*]; [Pro69His]compound heterozygous0.005873; 0.0003807P; LPDiagnosis establishedAbbreviations: AD, autosomal dominant; AR, autosomal recessive; F, Female; LP, Likely pathogenic; M, Male; P, Pathogenic; VUS, Variant of uncertain significance; y, years. a Reference transcript accession numbers and versions correspond to those used in the original targeted panel design and annotation. b Allele frequency of East Asia from GnomAD_EAS (gnomAD v2.1.1, http://gnomad‐sg.org/). Additional findings identified by whole‐exome sequencing in panel‐negative cases. compound heterozygous Abbreviations: AD, autosomal dominant; AR, autosomal recessive; F, Female; LP, Likely pathogenic; M, Male; P, Pathogenic; VUS, Variant of uncertain significance; y, years. Reference transcript accession numbers and versions correspond to those used in the original targeted panel design and annotation. Allele frequency of East Asia from GnomAD_EAS (gnomAD v2.1.1, http://gnomad‐sg.org/). Taken together, the molecular findings identified by the tiered sequencing strategy involved six genes (Figure 2). Initial hotspot screening identified findings in MT‐RNR1, GJB2, and SLC26A4, whereas WES contributed additional findings in ILDR1, CDH23, and SLC12A2. Distribution of molecular diagnoses and candidate findings identified by the tiered testing strategy. Bars represent the number of patients with molecular diagnoses or candidate findings involving each gene and are stacked according to genotype. The left group summarizes molecular diagnoses established by targeted hotspot screening, whereas the right group shows additional molecular diagnoses and a candidate finding identified by whole‐exome sequencing. The SLC12A2 variant was classified as a variant of uncertain significance (VUS) and was not considered a definitive molecular diagnosis. Abbreviations: Hom, homozygous; compound het., compound heterozygous; VUS, variant of uncertain significance. Characterization of Rare and Notable Variants SLC12A2 c.2936A>G (p. Glu979Gly) In case P7, a previously unreported missense variant SLC12A2 (c.2936A>G, p.Glu979Gly) was identified. According to the ACMG/AMP guidelines, the variant was classified as a VUS based on currently available evidence. P7 had congenital hearing loss, which overlaps with previous reports of DFNA78. The proband reported hearing impairment in her father, spouse, and son. This family history was considered compatible with a presumed autosomal dominant inheritance pattern (Figure 3A). However, only the proband underwent genetic testing. Segregation of the variant could therefore not be assessed. In silico prediction yielded discordant results across multiple computational algorithms. Multiple sequence alignment showed that the glutamic acid residue at position 979 of the SLC12A2‐encoded protein is highly conserved among mammalian species (Figure 4). Therefore, the clinical significance of this variant remains uncertain. Pedigrees and available genotypes of the P7 and P47 families. (A) Pedigree of the P7 family. The proband carried the previously unreported SLC12A2 c.2936A>G variant. Only the proband underwent genetic testing; therefore, segregation analysis could not be performed. (B) Pedigree of the P47 family. The proband carried two heterozygous CDH23 variants, c.4562A>G and c.5311C>T. The unaffected mother carried c.5311C>T but not c.4562A>G; therefore, the two variants were presumed to be in trans. Filled symbols indicate individuals with hearing loss, open symbols indicate individuals without reported hearing loss, arrows indicate the probands, and diagonal slashes indicate deceased individuals. Evolutionary conservation analysis of the SLC12A2 p. Glu979Gly variant. Multiple sequence alignment of SLC12A2 orthologs across representative mammalian species demonstrates that the Glu979 residue is highly conserved. The red box and arrow indicate the position of the affected residue. CDH23 c.5311C>T (p.Arg1771Ter) In case P47, two heterozygous CDH23 variants were identified, including a nonsense variant c.5311C>T (p. Arg1771Ter) and a missense variant c.4562A>G (p. Asn1521Ser). The truncating variant c.5311C>T, although recorded in public databases, has not been reported in Han Chinese patients with hearing loss to our knowledge. Clinically, the proband presented with congenital deafness and had an affected older sister, while other family members (parents, two sons, one daughter, and one grandson) had normal hearing. Neither the proband nor her affected sister reported visual symptoms or known ocular abnormalities. Genetic testing showed that the unaffected mother carried c.5311C>T but not c.4562A>G. The two variants were therefore presumed to be in trans (Figure 3B). Discussion This study applied a tiered genetic testing strategy combining targeted hotspot screening and WES in a Han Chinese cohort with hearing loss from Southwest China. The initial hotspot screening identified variants mainly in MT‐RNR1, GJB2, and SLC26A4. In addition, second‐tier WES identified additional rare variants beyond the hotspot panel. These findings included additional molecular diagnoses, a candidate finding, and an unresolved heterozygous finding. WES therefore broadened the observed variant spectrum, although not all additional findings established a molecular diagnosis. These findings illustrate the complementary roles of hotspot screening and WES in extending variant detection within a tiered testing strategy. In this cohort, MT‐RNR1 variants, particularly m.1555A>G, represented the largest component of the detected variant spectrum, followed by GJB2 and SLC26A4. The m.1555A>G variant was also the most frequent genotype among patients with an established molecular diagnosis. This distribution differed from that reported in a large Chinese hearing‐loss cohort, in which SLC26A4 and GJB2 were the leading molecular causes, whereas MT‐RNR1 accounted for a smaller proportion of diagnosed cases (Wu et al. 2022). The predominance of MT‐RNR1 may partly reflect the characteristics of this cohort. Participants were adults recruited from a single workplace, and a substantial proportion reported ototoxic drug exposure (Table 1). This exposure history is relevant because MT‐RNR1 variants are well‐established risk factors for aminoglycoside‐induced hearing loss (Prezant et al. 1993). The observed distribution may therefore be cohort‐specific rather than representative of the general Han Chinese population. WES established additional molecular diagnoses in P37 and P47. P37 carried two heterozygous ILDR1 variants, c.772C>T (p.Gln258Ter) and c.206C>A (p.Pro69His), and had hearing loss first noted at 5–6 years of age. Although ILDR1‐related hearing loss is commonly congenital or prelingual, childhood onset at approximately 5 years has also been reported (An et al. 2019). Thus, the age at onset is within the reported phenotypic spectrum. In patient P47, two heterozygous CDH23 variants were identified and were presumed to be in trans based on the maternal testing results. CDH23 encodes a calcium‐dependent cell–cell adhesion glycoprotein that plays a critical role in sensory signal transduction, particularly in the auditory system (GeneCards, n.d.), and has been implicated in Usher syndrome Type 1 as well as autosomal recessive nonsyndromic hearing loss. Variants in CDH23 can result in a broad phenotypic spectrum, ranging from congenital profound to progressive high‐frequency hearing loss (de Guimaraes et al. 2024; Usami et al. 2022). P47 had congenital hearing loss and an affected older sister. Neither reported visual symptoms nor known ocular abnormalities. The missense variant CDH23 c.4562A>G (p.Asn1521Ser) has been previously reported in individuals with hearing loss from European, African, and West Asian populations (Sloan‐Heggen et al. 2015; Morgan et al. 2018; Bazazzadegan et al. 2019; Abu Rayyan et al. 2020; Wonkam et al. 2022). In contrast, the nonsense variant c.5311C>T (p.Arg1771Ter) introduces a premature termination codon predicted to trigger nonsense‐mediated mRNA decay and loss of function. Loss‐of‐function CDH23 variants are known to be pathogenic (Bolz et al. 2001; Schultz et al. 2011). This variant was absent in East Asian populations in gnomAD and has not been reported in Han Chinese patients with hearing loss, to our knowledge. It was previously described in an Israeli pedigree with Usher syndrome (Sharon et al. 2020), and later identified in three unrelated Italian cases of inherited retinal disease (Karali et al. 2022). Our observation therefore provides additional population‐specific documentation of this variant in a patient with hearing loss. We also identified a previously unreported SLC12A2 variant, c.2936A>G (p.Glu979Gly), which was classified as a VUS according to the ACMG/AMP criteria. It was therefore retained as a candidate finding rather than an established molecular diagnosis. SLC12A2 encodes a membrane‐bound Na+/Cl− cotransporter that critically regulates ionic homeostasis and cellular volume maintenance in the cochlea. Mutai et al. (2020) proposed that SLC12A2 contributes to endolymph homeostasis by recycling K+ from the perilymph to the stria vascularis and reported a de novo variant affecting the same residue, c.2935G>A (p.Glu979Lys), in a Japanese infant with severe sensorineural hearing loss. Similarly, McNeill et al. (2020) subsequently described two familial cases of bilateral SNHL harboring the same p.Glu979Lys substitution. P7 had congenital hearing loss, which overlaps with the reported phenotype of DFNA78. The family history was compatible with a presumed autosomal dominant inheritance pattern. However, only the proband underwent genetic testing, and segregation of the variant could not be assessed. Detailed comparison of hearing severity and vestibular involvement was also limited by the available clinical information. The recurrence of hearing loss associated variants affecting codon 979, together with the conservation of Glu979, suggests that this residue may represent a functionally important site of the protein. Nevertheless, the evidence remains insufficient to establish a causal role for p.Glu979Gly. Additional segregation studies and functional assays will be required to clarify its clinical significance. Although WES was performed only in the 11 patients in whom no variants were detected by the initial hotspot screening, it enabled the identification of additional rare variants not covered by the hotspot panel, including additional molecular diagnoses and a candidate finding. This highlights its complementary role in investigating genetically heterogeneous cases. In settings where broad sequencing cannot be applied to all patients, such a stepwise strategy may represent a practical approach for balancing testing breadth with resource use. Several limitations should be acknowledged. First, the sample size was relatively limited, and whole‐exome sequencing was performed only in the 11 patients in whom no variants were detected by the initial hotspot screening. Patients with a single heterozygous hotspot variant were not further examined by WES. Additional pathogenic alleles may therefore have remained undetected. Second, family‐based genetic data were not available for all cases, restricting further segregation analysis and phase assessment. Third, due to the retrospective design, phenotypic information was not uniformly available across all participants. Clinical information was based mainly on interviews and available medical records. Detailed audiological data were not uniformly available, and standardized vestibular and ophthalmological assessments were not routinely performed. These limitations mainly affect the interpretation of individual cases. Future multicenter studies with comprehensive family‐based analyses and standardized clinical assessments will help improve variant interpretation and genotype–phenotype correlations. Conclusions In conclusion, this study expands the mutational spectrum of hereditary hearing loss in a Han Chinese cohort from Southwest China, including a previously unreported SLC12A2 candidate variant currently classified as a VUS and a CDH23 variant not previously reported in Han Chinese patients with hearing loss. These findings provide additional population‐specific genetic evidence and highlight the practical value of a tiered testing strategy integrating targeted hotspot screening with whole‐exome sequencing. Author Contributions Shimei Yuan: investigation, formal analysis, data curation, writing – original draft, writing – review and editing. Lanlan Xu: investigation. Tian Li: investigation. Yu Zhang: formal analysis, data curation, writing – original draft, writing – review and editing, visualization, investigation, validation. Yafang Wan: investigation, data curation. Changchun Niu: conceptualization, methodology, funding acquisition, supervision, writing – review and editing, project administration, validation. Wei Yang: conceptualization, methodology, resources, project administration, supervision, writing – review and editing. Funding This work was supported by the General Program of National Natural Science Foundation of China (82673032), the Chongqing special project for technological innovation and application development (CSTB2023TIAD‐KPX0063), and the Joint project of Chongqing Health Commission and Science and Technology Bureau (2024MSXM105). Ethics Statement This study was reviewed and approved by the Ethics Committee of Chongqing General Hospital, Chongqing University (Approval No. KY S2022‐030‐1). All procedures were conducted in accordance with institutional guidelines and the Declaration of Helsinki. Written informed consent was obtained from all participants prior to sample collection and genetic testing. Consent The authors have nothing to report. Conflicts of Interest The authors declare no conflicts of interest. Supporting information Table S1: Genes and variants included in the targeted hotspot panel. Table S2: Patients with variant findings that did not establish a definitive molecular diagnosis. Contributor Information Wei Yang, Email: yw74681179@126.com. Changchun Niu, Email: bright_star2000@163.com. Data Availability Statement The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions. References Associated Data Supplementary Materials Table S1: Genes and variants included in the targeted hotspot panel. Table S2: Patients with variant findings that did not establish a definitive molecular diagnosis. Data Availability Statement The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.