2012/02/09 Revision of MS. 1504994795618680 Dear Editor:
All authors are grateful for this opportunity to revise the paper. We also very much appreciate the valuable comments from all reviewers. As suggested, we have revised the manuscript and respond to the reviewers’ comments point by point as in the attached pages. As a result, the revised manuscript has been modified to accommodate all the reviewers’ concern and suggestions. We believe that we are able to fully address the questions raised by the reviewers. The revisions were highlighted with RED color in the revised text and tables. Thank you very much for your editorial effort and consideration.
Jim Jinn-Chyuan Sheu, PhD Associate Professor
Human Genetic Center
Department of Medical Research China Medical University Hospital TEL: +886-4-22052121 ext.2037 FAX: +886-4-22053366
Reviewer: Kun Huang
This paper focuses on one gene and tries to assess if the SNPs on this gene (Mucin 2) is associated with endometriosis and infertility. The technique and analysis is straight forward. But the collection of samples requires certain work (several hundred blood samples). There are a few SNPs shown to have relatively significant associations with the disease conditions.
1. However, one issue is that the p-values are not strong. If multiple test compensation (e.g., Bonferroni, the authors may also consider other approaches) is applied, these may not pass the threshold. The authors should address this issue.
Response: Thanks for your thoughtful question. It is true that the p-values for individual associations are not very strong (p > 0.01) (Table 1). However, multiple test compensation may be more suitable for considering multiple statistical comparison tests as one set to determine whether any differences between two groups exist. While in this case, instead of trying to distinguish the case and control group themselves, we aimed to find whether each individual SNP had different distributions in the case group vs. the control group. Thus, each comparison was rather independent to draw a conclusion. In addition, although Bonferroni or standard-Šidák correction provides a relatively straight-forward way in elimination of false-positives in a multi-test situation, it is at the expense of over-correction, especially for testing SNPs in LD (e.g. SNPs closely related as in this study of SNPs within one gene). A spectral-decomposition based method (Nyholt 2004) or permutation-test based correction may be a better alternative in adjusting for our multi-test situation, which would specify thresholds of P≤0.018 and P≤0.022, respectively, in order to keep the type 1 error rate at 5% for this study. If such thresholds were applied, only rs10902088 would be considered significant. To clarify this, we have added the description that p-values are calculated without multiple test corrections within footnote of Table 1.
2. The interesting thing is the combination of the three SNPs and the haplotype (T-C-A) shows stronger p-value in infertility. So what is the implication?
Response: Thanks for your incisive question. We also felt that the haplotype association with infertility is a very interesting result. In the individual SNP and infertility association test, rs10794288 and rs10902088 both were found to be significant. Consequently, combination of these two SNP showed a stronger association with infertility than each SNP alone (p = 0.0091) (Table 3). However, minor alleles of rs7103978 were much less common compared with the former two SNPs (supplementary table 1). Therefore, the insignificant p-value may due to our limited number of infertile patients (majority of endometriosis patients are still fertile) accompanied by the low minor allele frequency of rs7103978 (only 6% in Han Chinese in Beijing). If we had a larger sample group, the infertility
association might be more accurate for this SNP. Still, the fact that the major allele A may be a risk factor of endometriosis (Table 1) is consistent with the result that the T-C-A haplotype is slightly more significant (p = 0.0066) compared to the T-C haplotype (Table 3). Notably, it is currently still under debate whether haplotype of multiple SNP can provide greater power than individual SNPs to discriminate causative mutation loci in general, even though this question has been examined by many previous theoretical and empirical studies (Martin et al. 2000; Akey et al. 2001; Morris and Kaplan 2002).
3. It seems that one of them (rs10902088) changes amino acid (Asn to Lys). Do the other two lie on any regulatory site (e.g., enhancers)?
Response: Thanks for your question. The other two are silent mutations, which also lie in exon region (supplementary table 1). No regulatory site at these positions has been reported so far. Nevertheless, these two SNP may influence the efficiency of codon usage, according to the codon usage database maintained by Kazusa DNA Res. Institution (http://www.kazusa.or.jp/codon/). The minor allele of rs10794288 slightly increases the codon frequency from 21.8‰ to 25.1‰; while the minor allele of rs7103978 decreases the codon frequency from 15.8‰ to 7.8‰. Therefore, the minor G allele of rs7103978 may be more capable of reducing MUC2 production by increasing the odds of premature translation termination. Since expression of MUC2 were reported more prominent in neoplastic human gastric epithelium and malignant cervix uteri lesions (Riethdorf et al. 2000; Babu et al. 2006), decreased MUC2 level may help reduce the invasiveness of endometrium cells, which would behave as a protective factor against this disease. However, possible codon effect of rs10794288, which may be subtle given the codon percentage mentioned above, is opposite to our observations. Significance of this SNP may arrive from its high linkage to rs10902088. We added discussion of these in the result and discussion section on page 13 and 14.
Reviewer: Victor Jin
This work conducted with 195 endometriosis patients and 196 healthy controls at China Medical University Hospital to study genetic association of mucin2 (MUC2) with the risk of endometriosis and endometriosis-related infertility. This case-control study was Genotyping of six SNPs within MUC2 gene using Taqman genotyping assay. They found Endometriosis patients exhibit significantly lower frequency of the rs10794288 C allele, the rs10902088 T allele and the rs7103978 G allele. Haplotype analysis of the endometriosis associated SNPs in MUC2 also showed significantly association between the most common haplotypes and endometriosis or endometriosis-related infertility. The study is well-designed and important to genetic research.
References:
Akey, J., Jin, L., et al. (2001). "Haplotypes vs single marker linkage disequilibrium tests: what do we gain?" Eur J Hum Genet 9(4): 291-300.
Babu, S. D., Jayanthi, V., et al. (2006). "Expression profile of mucins (MUC2, MUC5AC and MUC6) in Helicobacter pylori infected pre-neoplastic and neoplastic human gastric epithelium." Mol Cancer 5: 10.
Martin, E. R., Lai, E. H., et al. (2000). "SNPing away at complex diseases: analysis of single-nucleotide polymorphisms around APOE in Alzheimer disease." Am J Hum Genet
67(2): 383-94.
Morris, R. W. and Kaplan, N. L. (2002). "On the advantage of haplotype analysis in the presence of multiple disease susceptibility alleles." Genet Epidemiol 23(3): 221-33. Nyholt, D. R. (2004). "A simple correction for multiple testing for single-nucleotide polymorphisms in linkage disequilibrium with each other." Am J Hum Genet 74(4): 765-9. Riethdorf, L., O'Connell, J. T., et al. (2000). "Differential expression of MUC2 and MUC5AC in benign and malignant glandular lesions of the cervix uteri." Virchows Arch
