A KCNQ4 c.546C>G Genetic Variant Associated with Late Onset Non-Syndromic Hearing Loss in a Taiwanese Population

Round 1

Reviewer 1 Report

The authors describe a mutation screening for potentially causative variants in 527 individuals who visited an ORL department. The main outcome is a relatively high frequency of one KCNQ4 variant, c.827G>C, in people with mild to profound hearing loss. The frequency of the variant is compared to persons with no known variant. In a PheWAS, the frequency of other diseases is compared between and non-carriers carriers of c827G>C.

Although the finding of a recurrent KCNQ4 variant associated with hearing loss in an interesting finding with potential diagnostic implications, the study contains several methodological flaws, the data analysis is not always appropriate and certain conclusions are not in line with the data.

Below I provide a non-exhaustive overview of some major concerns on this study.

• In table 1, an overview of the variants is provided. The allele frequency in TCVGH and TWB is provided, but not from GnomAD, which is the gold standard for the frequency of variants. GnomAD includes MAF for ethnicities all over the world, including east Asia.
• The c827G>C variant was found previously by Su et al. but no details are given about this publication - this reference is provided is an almost sneaky way. What was previously known about the c827G>C variant? This should be discussed in more detail.
• There is no discussion about the pathogenicity of the c827G>C variant. What is the CAT score? What is the pathogenicity according to the ACMG guidelines? Nowhere in the manuscript the authors mention that c827G>C causes a protein change (Phe182Leu). How does the variant relate to other variants in KCNQ4?
• The authors compare certain clincal characteristics (frequency of hearing loss, split according to severity) between the carriers of the c827G>C variant and 160 persons where no known mutation. These latter persons are referred to as “normal controls”, but this designition is not correct. They were included because they visited the ENT clinic and an audiogram was available – not a random sample of the population with normal hearing. It is not clear which research question is answered by comparing het clinical characteristics of the c827G>C to the individuals without a known mutations. If the aim is to check if the c827G>C carriers are a separate group with a distinctive audiongram, a principal component analysis or discriminant analysis may be a better option than comparing the frequency of mild, moderate and severe hearing loss between the two groups.
• In table 2, 3^rd There arer 134 persons with sensorineural/mixed hearing loss in the so-called notmal control group, but the sum of mild-to-profound is 68+51+15+5=138.
• The association between hearing loss and the c827G>C variant is tested by logistic regression, with age and gender as a covariate. Including age as a main effect implicitly assumes that the effect of age on the outcome (HL) is the same in carriers and non-carriers. This is a wrong assupmtion – table 3 shows that c827G>C causes late-onset hearing loss. The most likely explanation why no young individuals are found with c827G>C, is that young carriers are asymptomatic. The correct model would be a logistic model with an interaction between c827G>C carriership and age, whereby age modifies the effect of the c827G>C variant.
• The estimated odds ratio of the logistic regression model is 4.02, and in the discussion the authors state that c827G>C variant is associated with a 4-fold increase in risk for HL. This is not the correct interpretation of an odds ratio – this is the defintion of a relative risk. When not correcting for age, the RR would be
• The PheWAS study tests for association between c827G>C carriersip and the presence of several clincal conditions. Polyneuropathy, aortic aneurism and limb fractires are listed as signifciant hits with p<0.001, but there is no correction for multiple hypothesis testing. When testing a large number of hypotheses, the type 1 error increases. Even in the absence of any association, p-value below 0.05 will occur by chance alone. A Bonferroni correcion would be too strict here – as not all medical conditions are independent – but there should at least be an evaluation of the significance by e.g. a QQ-plot or a false discovery rate analysis.

These are fundamental issues that question the validity of the conclusions. I’m afraid a quick-fix for these issues is not possible. I would recommend the authors should revise their manuscript with the help of an expert in medical genetics and/or genetic epidemiology.

Author Response

Author’s Response to reviewer’s comments

Reviewer #1

Comments and Suggestions for Authors

The authors describe a mutation screening for potentially causative variants in 527 individuals who visited an ORL department. The main outcome is a relatively high frequency of one KCNQ4 variant, c.827G>C, in people with mild to profound hearing loss. The frequency of the variant is compared to persons with no known variant. In a PheWAS, the frequency of other diseases is compared between and non-carriers carriers of c827G>C.

Although the finding of a recurrent KCNQ4 variant associated with hearing loss in an interesting finding with potential diagnostic implications, the study contains several methodological flaws, the data analysis is not always appropriate and certain conclusions are not in line with the data.

Below I provide a non-exhaustive overview of some major concerns on this study.

• In table 1, an overview of the variants is provided. The allele frequency in TCVGH and TWB is provided, but not from GnomAD, which is the gold standard for the frequency of variants. GnomAD includes MAF for ethnicities all over the world, including east Asia.

Author’s response

As suggested by the reviewer, we have provided allele frequency of East Asia from GnomAD and modified the Table 1, in the revised manuscript.

• The c827G>C variant was found previously by Su et al. but no details are given about this publication - this reference is provided is an almost sneaky way. What was previously known about the c827G>C variant? This should be discussed in more detail.

Author’s response

Appreciate the reviewer’s suggestion, the KCNQ4 c.546C>G variant was found previously by Su et al. and we have addressed this issue in the Discussion section in the revised manuscript. Please see the revised manuscript for details(page 10, lines 260-283).

Reference:

- Su, C.C.; Li S.Y.; Yang J.J.; Su M. C.; Lin M.J. Studies of the effect of ionomycin on the KCNQ4 channel expressed in Xenopus oocytes. Biochem Biophys Res Commun.2006, 348, 295-300.

-Su, C.C.; Yang J.J.; Shieh J.C.; Su M. C.; Li S.Y. Identification of novel mutations in the KCNQ4gene of patients with nonsyndromic deafness from Taiwan. Audiol Neurootol.2007, 12, 20-26.

• There is no discussion about the pathogenicity of the c827G>C variant. What is the CAT score? What is the pathogenicity according to the ACMG guidelines? Nowhere in the manuscript the authors mention that c827G>C causes a protein change (Phe182Leu). How does the variant relate to other variants in KCNQ4?

Author’s response

Thank you for the valuable input. To judge the pathogenicity of thec.546C>G variant, CADD scores are taken from the CADD project web site (version v1.5). The value of CADD Raw score is 2.205. The positive value indicated that the KCNQ4 c.546C>G variant is likely to be simulated and more likely to have deleterious effect based on different genomic features. And CADD Phred-scaled score of 20.9 means that the variant is potentially pathogenic variants with cutoff value is set on 15.According to the ACMG guidelines and Annovar database, rs80358273 was classified as 2PM (pathogenic moderate) and 1PP (pathogenic supporting) evidence. Meanwhile, our report demonstrated a significant increased risk in patients with KCNQ4 c.546C>G variant. Although the risk did not reach a statistical significance, we did observe SNHL inmultiple unrelated patients with c.546C>G variant. Therefore, our study provided another evidence of moderate pathogenicity. Taken together, we judged the KCNQ4 c.546C>G as3PM and 1PP, ie. likely pathogenic by ACMG guidelines(pages10-11, lines 284-296).

Nowhere in the manuscript the authors mention that c827G>C causes a protein change (Phe182Leu).

Author’s response

Mutation F182L of KCNQ4 gene was investigated in the sub-cellular localization of HeLa cell, which resulted in proteins with impaired trafficking and localized in perinuclear vesicles, not localized to the membrane. Mutation F182L of KCNQ4 gene was not affect the protein structure of KCNQ4, but it affected the protein function of KCNQ4 and influence the function of potassium channel(page 10, lines 260-271).

How does the variant relate to other variants in KCNQ4?

Author’s response

In Taiwan, two silent mutations of uncertain clinical significance were reported. Variants like KCNQ4 c.648C>T causes a protein change (R216R) and KCNQ4 c.1503C>A causes a protein change (T501T) also have onset of symptoms during childhood. The animal study performed by Suet al. [18,22] was investigated that silent mutation-R216R did not alter the content of amino acid residue, but the neural network prediction system revealed that it can potentially create a novel splice donor site during transcription, and lead to loss of 66 nucleotides, might affect the protein structure of KCNQ4 and function of the K+ channel. In contrast, T501T mutant resulted in a protein that was localized to the membrane, like normal KCNQ4 protein. Therefore, mutation T501T of KCNQ4 gene did not affect the synthesis of KCNQ4 proteins nor their transportation to the cell membrane to form functional potassium channels. The role of T501T mutant in hereditary hearing loss need to be further study to evaluate the function in DFNA2/KCNQ4 gene(page10, lines 272-283).

• The authors compare certain clinical characteristics (frequency of hearing loss, split according to severity) between the carriers of the c827G>C variant and 160 persons where no known mutation. These latter persons are referred to as “normal controls”, but this designition is not correct. They were included because they visited the ENT clinic and an audiogram was available – not a random sample of the population with normal hearing. It is not clear which research question is answered by comparing het clinical characteristics of the c827G>C to the individuals without a known mutations. If the aim is to check if the c827G>C carriers are a separate group with a distinctive audiongram, a principal component analysis or discriminant analysis may be a better option than comparing the frequency of mild, moderate and severe hearing loss between the two groups.

Author’s response

We thank the reviewer for the correction. Taking account of the comment, we revised our case-control study design following the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline. In Table 2, we compared case group with SNHL/MHL to control group without SNHL/MHL. In this study, we aimed to evaluate the association ofKCNQ4 c.546C>G variant and non-syndromic hearing loss. We observed a significant difference of KCNQ4 c.546C>G carriership between 2 groups. The KCNQ4 c.546C>G variant was defined as carriers and KCNQ4 wild type was defined as non-carriers. We revised our manuscript accordingly(page3, lines 99-113).

• In table 2, 3^rdThere arer 134 persons with sensorineural/mixed hearing loss in the so-called notmal control group, but the sum of mild-to-profound is 68+51+15+5=138.

Author’s response

Thank you for bringing up this question. There were 134 patients with sensorineural/mixed hearing loss, including 65 mild hearing loss, 50 moderate hearing loss, 14 severe hearing loss, and 5 profound hearing loss. In addition, 4 patients were classified as sensorineural/mixed hearing loss, including 3 with mild hearing loss and 1 with moderate hearing loss, respectively.

• The association between hearing loss and the c827G>C variant is tested by logistic regression, with age and gender as a covariate. Including age as a main effect implicitly assumes that the effect of age on the outcome (HL) is the same in carriers and non-carriers. This is a wrong assupmtion – table 3 shows that c827G>C causes late-onset hearing loss. The most likely explanation why no young individuals are found with c827G>C, is that young carriers are asymptomatic. The correct model would be a logistic model with an interaction between c827G>C carriership and age, whereby age modifies the effect of the c827G>C

Author’s response

As indicated by the reviewer, this case-control study aimed to evaluate the association of KCNQ4 c.546C>G variant and non-syndromic hearing loss. Because the majority (11 out of 12) of KCNQ4 c.546C>G variant carriers aged over 70, we revised the Table 4 to include only participants aged over 70 to minimize the effect age to hearing loss. The association of KCNQ4 c.546C>G variant and sensorineural/mixed hearing loss was evaluated by univariable logistic regression.

• The estimated odds ratio of the logistic regression model is 4.02, and in the discussion the authors state that c827G>C variant is associated with a 4-fold increase in risk for HL. This is not the correct interpretation of an odds ratio – this is the definition of a relative risk. When not correcting for age, the RR would be

Author’s response

Odds ratio (OR) is used as an important metric of comparison of two or more groups. For rare events, the OR, given by [p₁/(1−p₁)]/[p₂/(1−p₂)], closely approximates the relative risk (RR), p₁/p₂. In this study, each subject has the similar length of observation time. In response to the comment, we removed the description of 4-fold increased risk in the manuscript.

• The PheWAS study tests for association between c827G>C carriersip and the presence of several clincal conditions. Polyneuropathy, aortic aneurism and limb fractires are listed as signifciant hits with p<0.001, but there is no correction for multiple hypothesis testing. When testing a large number of hypotheses, the type 1 error increases. Even in the absence of any association, p-value below 0.05 will occur by chance alone. A Bonferroni correcion would be too strict here – as not all medical conditions are independent – but there should at least be an evaluation of the significance by e.g. a QQ-plot or a false discovery rate analysis.

Author’s response

Thanks for the reviewer’s comments. The Bonferroni correction could be overly conservative because of the assumption that all tests are independent and that all phenotypes were independent, hence, not accounting for the correlation between the genetic variants due to linkage disequilibrium or phenotypes. Therefore, we evaluated the significance by QQ-plot of PheWAS in Figure 2 to evaluate whether the observed distribution was different from what would be expected under the null hypothesis. In Figure 2, QQ-plot of expected and observed p-values of the associations between SNPs and diseases showed clear enrichment of p-values. Thus, we could consequently expect our data to be suitable for the PheWAS of the KCNQ4 c827G>C, and we had sufficient power to detect modest association between this variant and diseases(page 9, lines 226-229; Figure 2).

These are fundamental issues that question the validity of the conclusions. I’m afraid a quick-fix for these issues is not possible. I would recommend the authors should revise their manuscript with the help of an expert in medical genetics and/or genetic epidemiology.

Author’s response

Thank you for your valuable suggestion. We modified our manuscript with the help of Dr. Ching-Heng Lin and Dr. Tzu-Hung Hsiao. Dr. Lin is an expert in epidemiology and public health. Dr. Hsiao is an experienced researcher in medical genetics. We thank them for the critical suggestion of this manuscript.

Author Response File: Author Response.pdf

Reviewer 2 Report

Paper is devoted to interesting topic and the abstract clearly reflects the investigation. Nevertheless, the proposed interpretation of the results is controversial. We can speak about a high frequency of the studied variant among elderly people with hearing loss, however statistics on the prevalence of this variant in a healthy population are not presented and the control group in this investigation consists of 160 people.

Comments:

1. The presentation of the material is crumpled. Line 82 is more about results.
2. Line 78-85: In the section on the materials and methods patient information is given very briefly. |The age of the patients is indicated only in the results. The characteristics of 32 thousand patients in the TPMI database are not given in the section, it is not clear which part of them has hearing loss and what other diseases, and how the patients got into this database one can only guess.
3. Line 77: The age of the participants in the total group of 32728 people is not clear.
4. The control group included persons under 60 years old, where they were taken from is not entirely clear.
5. Line 133-135: Perhaps this information should be in the section of materials and methods.
6. Line 135-136: Did we understand correctly that we are talking about variant carriers and non-carriers.
7. Line 139: Table 1 is a reductant for the material of this article, it is too detailed and is practically not described in the text, whether there is a need for it.
8. In Table 1 it is not clear what a heterozygous genotype is, whether there are compound heterozygotes among them.
9. Line 139 (table 1)  - is it correct to define the genotype with mitochondrial mutations as homozygous (4 last cases 26 CHR) – how correct is it to attribute mitochondrial mutations to the homozygous genotype?
10. Line 141: Table 2 does not give an idea of the association of hearing loss with other diseases.
11. Line 158: Fig ! is very briefly commented in the text, the age of the patients is not indicated.
12. Line 160: Contents of Fig 2 are commented very briefly.
13. Is it possible to interpret these results as an increase in risk associated with this genetic variant, which was noted in all the patients presented. Why these states are not discussed

Author Response

Reviewer#2

Comments:

1. The presentation of the material is crumpled. Line 82 is more about results.

Author’s response

We appreciate for your valuable suggestion, and we have modified the sentence in the Methods section (page 3, lines 99-113). As illustrated in Supplementary Figure 1, we selected participants who has audiograms recorded from otorhinolaryngology in TCVGH during the period from June 2019 to June 2020. We extracted participants aged>18 years with available Pure Tone Audiometry results. There are 527 participants with PTA in the data source, and 179 were excluded without records. The final study population included 348 patients with available audiograms, and there were 236 persons with sensorineural/mixed hearing loss (MHL).Of these, 112 participants without SNHL/MHL, 5 were excluded for mild conductive hearing loss (CHL), hence, there were 107 participants without KCNQ4 c.546C>G variant and with audiograms were selected as the control group.

2. Line 78-85: In the section on the materials and methods patient information is given very briefly. |The age of the patients is indicated only in the results. The characteristics of 32 thousand patients in the TPMI database are not given in the section, it is not clear which part of them has hearing loss and what other diseases, and how the patients got into this database one can only guess.

Author’s response

As suggested by the reviewer, we provided the information of how we recruited our participants in the Methods section (pages 2-3, lines 85-97).

This case-control study was conducted from the Taiwan Precision Medicine Initiative (TPMI).Between June 2019 and June 2020, TPMI identified 32,728 patients aged >18 years(15,249 men and 17,479 women), who recruit from 28 medical and surgical outpatient clinics from Taichung Veterans General Hospital (TCVGH) with genotyping information, demographics, medical history, as well as biochemical reports, and all of the participants provided informed consent. As illustrated in Supplementary Figure 1, we selected participants who has audiograms recorded from the otorhinolaryngology outpatient clinics in TCVGH TPMI cohort during the period from June 2019 to June 2020. In total, there are 527 participants with Pure Tone Audiometry (PTA) test result in the data source, and 179 were excluded due to lack of records. The final study population included 348 patients with available audiograms.

3. Line 77: The age of the participants in the total group of 32728 people is not clear.

Author’s response

Appreciate the reviewer’s comment and we provided this information in the Methods section (page 3, lines 91).

Our study cohort consisted of 32,728 patients (mean age 57.3 ± 15.1 years, 15249 males, 17479 female) with genotyping information, demographics, medical history, as well as biochemical reports, and all of the participants provided informed consent.

4. The control group included persons under 60 years old, where they were taken from is not entirely clear.

Author’s response

Thank you for the suggestion.

The final study population included 348 patients with available audiograms; of them, 236 persons with sensorineural/mixed hearing loss (MHL) were the case group. Among 112 participants without SNHL/MHL, five were excluded for mild conductive hearing loss (CHL), hence, 107 participants without KCNQ4 c.546C>G variant and with audiograms were selected as the control group. Finally, we analyzed the association between KCNQ4 c.546C>G variant and non-syndromic hearing loss. In the SNHL/MHL group, 12 patients who had the KCNQ4 c.546C>G variant were defined as KCNQ4 c.546C>G variant carriers, and 224 patients who without KCNQ4 c.546C>G variant were defined as KCNQ4 c.546C>G variant non-carriers. We revised manuscript(page 3, lines 99-113).

5. Line 133-135: Perhaps this information should be in the section of materials and methods.

Author’s response

As indicated by the reviewer, this sentence has been modified and this information has been provided in the Methods section.

6. Line 135-136: Did we understand correctly that we are talking about variant carriers and non-carriers.

Author’s response

Yes, it refers to variant carriers and non-carriers. It has been rectified in the revised manuscript.

7. Line 139: Table 1 is a reductant for the material of this article, it is too detailed and is practically not described in the text, whether there is a need for it.

Author’s response

Thank you for the comments. We added the description of hearing loss related genetic variants in the result section. In Table 1, our study indicated that the minor allele frequency (MAF) of KCNQ4 c.546C>G was 0.61% at TCVGH, higher than the MAF of East Asian population in GnomAD database. This result suggested that KCNQ4 c.546C>G was the most common one among the autosomal dominant nonsyndromic hearing loss in a Taiwanese population(pages 4-5, lines 159-171).

8. In Table 1 it is not clear what a heterozygous genotype is, whether there are compound heterozygotes among them.

Author’s response

A heterozygous genotype stands in contrast to a homozygous genotype. Heterozygous genotype refers to having inherited different forms of a particular gene from each parent.

Heterozygous genotype of KCNQ4 c.546C>G was CG, there are 398 participants have CG genotype, and these alleles are located at same loci at chr1:40818518. There are no compound heterozygotes among them. We revised our manuscript(page 4-5, lines 163-169).

9. Line 139 (table 1)  - is it correct to define the genotype with mitochondrial mutations as homozygous (4 last cases 26 CHR) – how correct is it to attribute mitochondrial mutations to the homozygous genotype?

Author’s response

Thank you for pointing this the Homoplasmic/heteroplasmic mitochondrial DNA mutation should be used.

In a previous study, the 12S rRNA gene in the mitochondrial DNA (mtDNA) has been shown to be a hot spot for non-syndromic sensorineural hearing loss, as four deafness-associated mtDNA mutations including A1555G, T1095C, A827G [Xing G et al., 2006], and mutations at position 14709 in the mitochondrial tRNA gene [Perucca-Lostanlen D et al., 2002]. Sequence analysis of the mitochondrial 12S rRNA and tRNA genes led to the identification of thehomoplasmicA1555G, T1095C, A827G, and TA827G mutation in all maternal relatives.

We revised the Table 1 accordingly.

- Xing G, Chen Z, Wei Q, Tian H, Li X, Zhou A, Bu X, Cao X. Maternally inherited non-syndromic hearing loss associated with mitochondrial 12S rRNA A827G mutation in a Chinese family. Biochem Biophys Res Commun. 2006 Jun 16;344(4):1253-7. doi: 10.1016/j.bbrc.2006.04.033. Epub 2006 May 2. PMID: 16650816.

- Perucca-Lostanlen D, Taylor RW, Narbonne H, Mousson de Camaret B, Hayes CM, Saunieres A, Paquis-Flucklinger V, Turnbull DM, Vialettes B, Desnuelle C. Molecular and functional effects of the T14709C point mutation in the mitochondrial DNA of a patient with maternally inherited diabetes and deafness. Biochim Biophys Acta. 2002 Dec 12;1588(3):210-6. doi: 10.1016/s0925-4439(02)00167-9. PMID: 12393175.

10. Line 141: Table 2 does not give an idea of the association of hearing loss with other diseases.

Author’s response

Thanks for the suggestion; we added the baseline characteristics and comorbidities of study populations in Table 2.

11. Line 158: Fig ! is very briefly commented in the text, the age of the patients is not indicated.

Author’s response

Appreciate the reviewer’s comment and we have modified the contents in the text and indicated the age of carriers in the Figure 1(page 8, lines 208-211).

12. Line 160: Contents of Fig 2 are commented very briefly.

Author’s response

As reviewer’s suggestions, and we have added the figure legends and detailed description of Figure 2 in the revised manuscript(pages 9-10, line 245-249).

13. Is it possible to interpret these results as an increase in risk associated with this genetic variant, which was noted in all the patients presented. Why these states are not discussed

Author’s response

Appreciate the reviewer’s suggestion, our results suggested that KCNQ4 c.546C>G variant is likely to be a potential pathogenic variant of ADNSHL in the Taiwanese population, and we have addressed this issue in the Discussion section(pages 10-11, lines 271-295, page 12, lines 339-342).

Author Response File: Author Response.pdf

Reviewer 3 Report

Manuscript ID: genes-1408258

Title: A KCNQ4 genetic variant associated with late onset non-syndromic hearing loss in a Taiwanese population

First Author: Ting-Ting Yen

The paper reports on association of a variant of KCNQ4 (c.546C>G) with post-lingual hearing loss in the elderly population in Taiwan. They studied the phenotype and clinical features of this variant in the patients representing hearing impairment by performing pure tone audiometry test and phenome-wide association (PheWAS) analysis. They showed that KCNQ4 c.546C>G is associated with sensorineural high frequency and progressive hearing loss. They also found an association between this variant and aortic aneurysm, fracture of lower limb and polyneuropathy in diabetes.

The study highlights the genotype-phenotype correlations in the patients presenting with age-related hearing loss and a specific variant of KCNQ4. This can help for genetic counseling and early intervention to prevent profound hearing impairment in these patients.

Comments:

-Page 1, line 28: Please replace the SNP name “rs80358273” with the variant name “c.546C>G”.

- Page 3, lines 109-110: ”… association between different phenotypes of variant c.546C>G and 3846". What does 3846 indicate?

- Page 3, line 112: "(Figure 2)” should be added at the end of the sentence.

- Page 3, line 137: Please add “Supplementary Figure 1” after variant: “… c.546C>G variant (Supplementary Figure 1)." The same line, the name of the variant should be written: “… patients with KCNQ4 c.546C>G variant …”.

- Page 4, Table 1: Please complete the abbreviations below the Table.  TWB, Taiwan Biobank; TCVGH, Taichung Veterans General Hospital should be added.

-Page 4, Table 1, column 1 (CHR): For mitochondrial genes, what does 26 indicate? The mitochondrial chromosome should be termed "chrM" (as termed in the UCSC Genome Browser, GRCh38 assembly).

- Page 5, line 146: Please add the name of the variant: “…group with KCNQ4 c.546C>G variant …”. Same in Page 7, line 165: “…plot for KCNQ4 c.546C>G variant …”. Please check it through the text.

- Page 5, Table 3: Please add “carriers” for KCNQ4 c.546C>G, and also replace “Normal control” with “Non-carrier controls”.

-Page 6, line 160: Please add the name of the gene “…PheWAS between KCNQ4 c.546C>G and …”

- Page 7, Section 3.3: This section should be rewritten with more explanation.

- Please place the figures legends below the figures.

- Figure 1: The audiograms are blurry but perhaps low resolution for reviewing purposes? The figure legend needs to be more detailed.

- Figure 2, page 6 and 7: The table apparently is a part of the figure 2? If so, please represent it correctly and give some explanations in the figure legend.

 -Table 2, line 141:  The name of the variant should be written in the title: “Association between KCNQ4 c.546C>G variant and hearing loss”.

Author Response

Reviewer#3

Comments:

-Page 1, line 28: Please replace the SNP name “rs80358273” with the variant name “c.546C>G”.

Author’s response

As reviewer’s suggestions, and we have changed “rs80358273” to “c.546C>G” in the Abstract section.

- Page 3, lines 109-110: ”… association between different phenotypes of variant c.546C>G and 3846". What does 3846 indicate?

Author’s response

Thank you so much for your careful check, and this typo was rectified.

- Page 3, line 112: "(Figure 2)” should be added at the end of the sentence.

Author’s response

As suggested by the reviewer,"(Figure 2)” had be added at the end of the sentence(page 4, line 139).

- Page 3, line 137: Please add “Supplementary Figure 1” after variant: “… c.546C>G variant (Supplementary Figure 1)." The same line, the name of the variant should be written: “… patients with KCNQ4 c.546C>G variant …”.

Author’s response

As suggested by the reviewer, we have added these words in the revised manuscript.

- Page 4, Table 1: Please complete the abbreviations below the Table.  TWB, Taiwan Biobank; TCVGH, Taichung Veterans General Hospital should be added.

Author’s response

As suggested by the reviewer, we have revised Table 1.

-Page 4, Table 1, column 1 (CHR): For mitochondrial genes, what does 26 indicate? The mitochondrial chromosome should be termed "chrM" (as termed in the UCSC Genome Browser, GRCh38 assembly).

Author’s response

Thank you so much for your careful check, and this mistake has been rectified.The mitochondrial chromosome had been termed "chrMT". "chrMT" is the revised Cambridge Reference Sequence mitochondrion (NC_012920).

- Page 5, line 146: Please add the name of the variant: “…group with KCNQ4 c.546C>G variant …”. Same in Page 7, line 165: “…plot for KCNQ4 c.546C>G variant …”. Please check it through the text.

Author’s response

As suggested by the reviewer, we have added the name of the variant throughout the revised manuscript.

- Page 5, Table 3: Please add “carriers” for KCNQ4 c.546C>G, and also replace “Normal control” with “Non-carrier controls”.

Author’s response

In this case-control study, the KCNQ4 c.546C>G variant was defined as carriers and KCNQ4 wild type was defined as non-carriers controls.

-Page 6, line 160: Please add the name of the gene “…PheWAS between KCNQ4 c.546C>G and …”

Author’s response

As suggested by the reviewer, we have added the name of the gene in title of the Figure 2.

- Page 7, Section 3.3: This section should be rewritten with more explanation.

Author’s response

Appreciate the reviewer's comment and we have provided more explanation in this section (page 9, lines 222-225).

- Please place the figures legends below the figures.

Author’s response

We have placed the figures legends below the figures in the revised manuscript.

- Figure 1: The audiograms are blurry but perhaps low resolution for reviewing purposes? The figure legend needs to be more detailed.

Author’s response

The resolution of Figure 1 resolution is 300 dpi (dots per inch) for image, and we have provided more information of the Figure 1 in the revised manuscript.

- Figure 2, page 6 and 7: The table apparently is a part of the figure 2? If so, please represent it correctly and give some explanations in the figure legend.

Author’s response

The table is a part of the Figure 2. In the revised manuscript, the table and Figure have been combined into one Figure as Figure 2. As reviewer’s suggestions, we have given more explanations in the figure legends of Figure 2.

 -Table 2, line 141:  The name of the variant should be written in the title: “Association between KCNQ4 c.546C>G variant and hearing loss”.

Author’s response

As suggested by the reviewer, we have added the name of the variant in the Table 3.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The authors have made a strong effort to improve the manuscript,  with the help of experts in the field.  They have appropriately replied to my previous remarks. I have no further comments.

Reviewer 2 Report

All comments have been corrected.