THE STUDY

1) The criteria of inclusion of patients in this study should be specified in more detail, and the number of cases with a clinical diagnosis of WS1 and WS2 should be reported as well. Taking into account Table 1, how could you include cases 6, 10 and 18 (n.a. in Downloaded from bmjopen.bmj.com on May 29, 2014 - Published by group.bmj.com all columns describing their phenotypes) in this study?

2) Mutations in MITF are not a major cause of WS2 (Introduction, paragraph 3, line 1), since it is estimated that they only account for 15% of the cases (see references 1 and 18). The contribution of SNAI2 is controversial (see comment on this point in reference 18), but almost irrelevant anyway. Also, the involvement of SOX10 in WS2 is not mentioned (see Bondurand et al., Am J Hum Genet 2007; 81:1169). It is estimated that SOX10 mutations are found in 15% of WS2 cases. The Introduction section should include these data.

3) More information is needed about the MLPA assay that was used. Were you using the MRC Holland kit? If so, it should be mentioned; if not, the method should be described in more detail.

RESULTS & CONCLUSIONS 1) Mutations in MITF only account for 15% of the WS2 cases. The contribution of SNAI2 is controversial, but almost irrelevant anyway. It is estimated that SOX10 mutations are found in 15% of WS2 cases. Given this background, it is surprising that the authors have found MITF mutations in all their WS2 cases. If really so, this point should be discussed in detail in the Discussion section. But, are only elucidated cases being reported?

2) Some data are just provided without any indication of supporting evidence (see also comment 3). Results, paragraph 1, last sentence. How did you prove that the two mutations were in the same chromosome? By segregation analysis? Please specify.

3) No data are presented that really support the de novo origin of the mutations that are reported in the manuscript. Was paternity investigated? I would not ask for a formal paternity test, but at least 6-7 polymorphic microsatellite markers from different chromosomes should be examined. REPORTING & ETHICS In the Materials and Methods section, Patients, nothing is said about Informed Consents and/or approval of the study by any Ethics Committee.

GENERAL COMMENTS This manuscript reports on 16 novel mutations in the PAX3 and MITF genes in cases of Waardenburg syndrome, which is a significant contribution to the spectra of mutations in these genes.

Major comments 1) In the Materials and Methods section, Patients, nothing is said about Informed Consents and/or approval of the study by any Ethics Committee.

2) The criteria of inclusion of patients in this study should be specified in more detail, and the number of cases with a clinical diagnosis of WS1 and WS2 should be reported as well. Taking into account Table 1, how could you include cases 6, 10 and 18 (n.a. in all columns describing their phenotypes) in this study?

3) Mutations in MITF are not a major cause of WS2 (Introduction, paragraph 3, line 1), since it is estimated that they only account for 15% of the cases (see references 1 and 18). The contribution of SNAI2 is controversial (see comment on this point in reference 18), but almost irrelevant anyway. Also, the involvement of SOX10 in WS2 is not mentioned (see Bondurand et al., Am J Hum Genet 2007; 81:1169). It is estimated that SOX10 mutations are found in 15% of WS2 cases. The Introduction section should include these data. Given this background, it is surprising that the authors have found MITF mutations in all their WS2 cases. If really so, this point should be discussed in detail in the Discussion section. But, are only elucidated cases being reported? This issue is related to major comment 2.

4) Some data are just provided without any indication of supporting evidence (see also major comment 5). Results, paragraph 1, last sentence. How did you prove that the two mutations were in the same chromosome? By segregation analysis? Please specify.

5) No data are presented that really support the de novo origin of the mutations that are reported in the manuscript. Was paternity investigated? I would not ask for a formal paternity test, but at least 6-7 polymorphic microsatellite markers from different chromosomes should be examined. Minor comments 1) Abstract, Objectives, Line 3 (and also in Page 2, Article Focus). Craniofacial dysmorphism is found only in WS1 and WS3, and so it is not a general feature of WS. 2) Abstract, Results. Last sentence is a comment, not a result, and it should be deleted or rephrased. 3) Abstract, Conclusion. The abbreviation "TS" should be avoided. 4) Introduction, paragraph 1, last line. Complete the last sentence "in addition to WS2", since WS4 does not include the craniofacial or limb malformations that are characteristic of WS1 or WS3. 5) More information is needed about the MLPA assay that was used. Were you using the MRC Holland kit? If so, it should be mentioned; if not, the method should be described in more detail. 6) Page 6, paragraph 3, line 4. Was the parent affected or not? 7) Page 6, paragraph 3, lines 6-7. Figures 1 and 2 do not summarize all previously published mutations, just their locations. 8) Nomenclature of mutations should be checked with the freely available Mutalyzer software. The names of the p.Arg37fs and p.Ser197fs mutations are incomplete. 9) The severity of the hearing impairment should be reported according to the standard rules (mild-moderate-severe-profound). What does (+) mean in Table 1, case 1? 10) Page 8, line 4. "All of these presented hearing" ?? 11) Page 8, last line. The sentence should be completed: "loss of protein function leading to haploinsufficiency seems to be the disease-causing mechanism for WS1". The mechanism is haploinsufficiency, not loss of function (one allele remains functional). 12) Some typos: caucasian (Caucasian; abstract, design and patients, line 1); fotographs (photographs; abstract, design and patients, line 5); german (German; abstract, setting, line 1); larger deletions (large deletions; page 2, key messages, and page 8, last paragraph, line 4).

The authors have complied with some of the requests in my previous referee's report, but some important comments have not been addressed, as follows.

1) In this class of study, it is not acceptable to present only cases who were positive in the molecular screening. The authors are not presenting a case, but a large cohort. The proportion of patients with a clinical diagnosis of WS who were confirmed (and those who were not) at the molecular level is meaningful. It allows comparison between cohorts of different studies, it allows to check the accuracy of the clinical criteria of inclusion, and it indicates how many cases remain unelucidated (who are candidate subjects for investigating the hypothetical implication of novel genes in the syndrome). I do not understand the bizarre reticence of the authors to present the whole data, since for sure they know how many cases they screened at the molecular level. Accordingly, it is also not acceptable to indicate "clinical designation" as a criterion for including cases 6 and 18. This "clinical designation" is for sure based on clinical findings that should be reported in Table 1.

2) In the legend to Table 1, for hypothetically de novo mutations, it should be added that "no paternity testing was performed". I recommend that "de novo?" should replace "de novo" for patients 8, 13 and 19 in Table 1.

3) Page 7, lines 2-3. I must insist on changing the sentence to say "The positions of these mutations and all previously published mutations are summarized...."

4) The authors have used the "short description" format of HGVS rules. I strongly recommend to use the "long description" format, which is much more informative, and can be checked by using the Mutalyzer software. Note that this also allows to detect incorrect names. For example, c.111dupC in PAX3 results in p.Val38Argfs*76. In its short form, it would be p.Val38fs (instead of p.Arg37fs).

5) Page 8, lines 5-6. "Loss of protein function respectively haploinsufficiency seems to be the disease causing mechanism for WS1". Please rephrase this sentence, it can be hardly understood.

6) Table 1. Reporting two sequence variants in the same allele should follow the HGVS rules: [c.28T>A; c.33+6del7] (patient 11).

The authors have modified the manuscript according to the criticisms in my previous reports. I have only a couple of minor comments:

1) I recommend to include a sentence in "Patients" explaining what the authors have replied to this referee about the composition of the cohort under study, i.e. that the 19 reported patients are those positive for mutation identification among over 120 cases with a clinical temptative diagnosis of Waardenburg syndrome, referred to their laboratory by many different clinicians.

2) Abstract, Results paragraph. "16" patients are mentioned here, but the right figure is 15.

3) Table 1. The "n.a." abbreviation should be explained.