Author's response to reviews Title: Genetic variations in APPL2 are associated with overweight and obesity in a Chinese population with normal glucose tolerance Authors:

2. Major Compulsory Revisions “APPL2, an isoform of APPL1 that forms a dimer with APPL1, function as a negative regulator of adiponectin signaling via interaction with adiponectin receptors in muscle cells.” This paragraph has very close resemblance to Wang et al, J Biol Chem. 2009 November 13; 284(46): 31608–31615. [APPL2, an isoform of APPL1 that forms a dimer with APPL1, can interacts with both AdipoR1 and AdipoR2 and acts as a negative regulator of adiponectin signaling in muscle cells.] Response: This paragraph was modified in the abstract section. We wrote, "APPL1 and APPL2 are two adaptor proteins, which can mediate adiponectin signaling via binding to N terminus of adiponectin receptors in muscle cells." Thank you. 3. Minor Compulsory Revisions Please use sex instead of gender throughout the manuscript. Gender is a social construct while sex is a biological construct. Response: We used "sex" instead of "gender" throughout the revised manuscript according to your suggestion. Thank you. 4. Major Compulsory Revisions BMI and WHR are usually associated and are co linear. Therefore testing for an association with each separately does not add any new information. Response: BMI, waist circumference and WHR are three most common anthropometric measurements to assess obesity in practice. Among the three, BMI, which is an internationally recommended parameter to classify obesity, can reflect generalized obesity. However, it ignores body fat distribution. Compared with BMI, waist circumference and WHR can describe central fat distribution and better reflect the accumulation of intra-abdominal fat. As intra-abdominal obesity rather than subcutaneous fat is closely related to metabolic risks, waist circumference and WHR are better predictors of the risks of such diseases. Therefore, we analysed the asssociations of SNPs with obesity-related measurements separately in the manuscript. (Feller S, et al. Body mass index, waist circumference, and the risk of type 2 diabetes mellitus: implications for routine clinical practice. Dtsch Arztebl Int 2010, 107(26):470-476.) (Cheng CH, et al. Waist-to-hip ratio is a better anthropometric index than body mass index for predicting the risk of type 2 diabetes in Taiwanese population. Nutr Res 2010, 30(9):585-593) (Schulze MB, et al.: Comparison of anthropometric characteristics in predicting the incidence of type 2 diabetes in the EPIC-Potsdam study. Diabetes Care 2006; 29(8): 1921–3.) Background 5. Minor Compulsory Revisions Obesity is already an epidemic. Please revise the introductory paragraph. Response: The paragraph in the background section has been revised. We wrote "In recent years, the worldwide prevalence of obesity has increased dramatically and has became a major global epidemic impacting on morbidity and mortality." Thank you. 6. Minor Compulsory Revisions “are two adaptor proteins, which binds to adiponectin receptors and mediate adiponectin signaling and function [4-6].” Please correct typographical and grammatical errors. Response: According to comment 2, this sentence was modified in the background section. We wrote "are two adaptor proteins, which can mediate adiponectin signaling via binding to N terminus of adiponectin receptors in muscle cells." Thank you. Methods Participants 7. Major Compulsory Revisions “according to the criteria set by WHO that classified all individuals into three groups: normal weight (BMI<25 kg/m2), overweight (25 BMI <30 kg/m2) and obese (BMI #30 kg/m2) [10].” Minor Compulsory Revisions: Please remove additional characters. Response: The mistake was modified in the revised manuscript. Thank you. 8. Major Compulsory Revisions In addition, the BMI cutoffs provided here are more appropriate to Caucasian populations. Several studies showed different cutoffs for Asian populations. Based on ROC curve, the appropriate BMI cutoff for Chinese adults is 24 and waist circumference of 80 cm (for e.g. see Wildman et al. American Journal of Clinical Nutrition, Vol. 80, No. 5, 1129-1136, November 2004). Response: In the revised method section, we modified the BMI cutoff for Chinese adults. We wrote, "BMI was used to assess generalized obesity according to the Chinese criteria that classified all individuals into three groups: normal weight (BMI < 24 kg/m 2 ), overweight (24 ≤ BMI < 28 kg/m 2 ) and obese (BMI ≥ 28 kg/m 2 )." (Bei-Fan Z: Predictive values of body mass index and waist circumference for risk factors of certain related diseases in Chinese adults: study on optimal cut-off points of body mass index and waist circumference in Chinese adults. Asia Pac J Clin Nutr 2002; 11(Suppl. 8):S685–S693). Accordingly, all the statistics were modified based on cutoff for Chinese adults. The results were similar to the previous and updated in the revised manuscript. 9. Major Compulsory Revisions A description of the study participants should be given in a simple demographic table to include age, sex distribution, BMI and WHR. It is also important to show how many individuals are in each group of BMI cutoffs. Response: We added a table (Table 1 Clinical characteristics of study population) in the revised manuscript to show the basic characteristics of our study population. Thank you. Single nucleotide polymorphism (SNP) selection and genotyping 10. Major Compulsory Revisions The SNP call rate <95% is low and signify genotyping errors. Response: We are very sorry we made a mistake in our calculation on SNP call rate in the manuscript. As we genotyped SNPs of APPL2 in 1,892 unrelated type 2 diabetic patients together with 1808 non-diabetic adults which all included in calculating, the SNPs call rates were inaccurate in previously submitted manuscript. The total individuals in our study were 1,806 non-diabetic adults (The explanation can be seen in the response to reviewer Francis Vasseur’s comment 6). Therefore, the SNP call rate was modified in the revised single nucleotide polymorphism (SNP) selection and genotyping section. We wrote "The call rates for rs2272495, rs10861360, rs1196768, rs3751191 and rs1107756 were 97.0%, 95.6%, 96.3%, 96.5% and 96.8%, respectively." The calculation details can be seen in the responses to comment 13 and 14. Statistical analysis 11. Major Compulsory Revisions Sample size and power calculations should be provided. Suppose: This was added into the “statistical analysis” section of the revised manuscript. We wrote, "We estimated study power using QUANTO. Assuming an additive model with the minor allele frequencies of 0.1, 0.2, 0.3 and 0.4 in our Chinese population, our sample size has 0.53, 0.77, 0.87 and 0.91 power to detect an OR of 1.25 at an α level of 0.05. " Results 12. Major Compulsory Revisions Table 1 overweight and obesity category should be revised with more appropriate cutoffs for this population. Response: We revised this table with more appropriate BMI cutoff for Chinese adults. And the results were updated in the revised manuscript. Thank you. 13. Major Compulsory Revisions According to this table there are 552 individuals [based on rs2272495 genotype count] with BMI>25 and 1200 with BMI<25. Total individuals=1752 [96.9% of total sample]. The SNP call rate was stated as 95.3%, were did the additional individuals come from? Response: According to the table 2 in the revised manuscript, there are 741 individuals [based on rs2272495 genotype count] with BMI ≥ 25 and 1011 with BMI < 25. Total individuals = 1752 [97.0% of total sample (1806)]. The SNP call rate for rs2272495 was stated as 97.0% in the revised manuscript. 14. Similarly other SNPs numbers do not add up. Please explain and account for all individuals. Response: There are 730 individuals [based on rs1107756 genotype count] with BMI ≥ 25 and 1018 with BMI < 25. Total individuals = 1748 [96.8% of total sample (1806)]. The SNP call rate for rs1107756 was stated as 96.8% in the revised manuscript. Similar to the rs2272495 and rs1107756, the call rates for rs10861360, rs1196768 and rs3751191were sated as 95.6% [1727/1806], 96.3% [1739/1806] and 96.5% [1743/1806]. 15. Major Compulsory Revisions In addition these numbers also show that 31.5% of the study population, described by authors as adults of Hans Chinese ancestry are overweight/obese using an already high cutoff of BMI (25). This rate is much higher than reported statistics on obesity in China. With lack of description of the study sample and unusual rate of obesity, the interpretation of the study results is speculative. Response: As age factor may influence the prevalence of obesity, the rate of overweight/obese in our manuscript was higher than reports on obesity in China. A cross-sectional study on prevalence of obesity in Chinese men showed the age-adjusted prevalence of general and centralized obesity among Chinese men living in urban Shanghai (Lee SA, et al. Prevalence of obesity and correla-tions with lifestyle and dietary factors in Chinese men. Obesity (Silver Spring) 2008; 16: 1440-7). This study was conducted in 61,582 Chinese men aged above 40. The average age and BMI were 54.9 years old and 23.7 kg/m 2 . The prevalence of overweight was 48.6% and obesity, 10.5%, respectively (BMI was used to measure overweight (23≤BMI<27.5) and obesity (BMI≥ 27.5) based on the WHO recommended criteria for Asians in this article). Our results were much similar to this study. Our study included 1,806 non-diabetic Chinese men aged above 40 years old. The average age and BMI were 57.3 years old and 23.6 kg/m 2 . The prevalence of overweight was 33.3% and obesity, 8.8% (the BMI cutoff for overweight and obesity in our study is 24 and 28, respectively, according to Chinese criteria). Table 1 in the revised manuscript described the basic characteristics of our study population. 16. The authors should also attempt to replicate these findings in an independent sample. Response: Thank you very much for your suggestion. Our findings need further confirmation in a larger independent sample. We added discussion on this point in the last paragraph of the revised discussion section. We wrote," Secondly, we found associations of rs2272495 and rs1107756 with overweight/obesity and obesity-related measurements without replicating our findings in a larger independent sample. Further replications in other cohorts are needed to confirm the influence of genetic variation within this locus on overweight/obesity or metabolic traits in the Chinese population." 17. Table 2 is not meaningful as all these measures are co linear and the authors are testing the same association of the SNP with adiposity multiple times to show the same results without correcting for these additional tests. Response: Compared with BMI, which reflects generalized obesity, waist circumference and WHR can describe central fat distribution and better reflect the accumulation of intra-abdominal fat. Therefore, we analysed the associations of SNPs with obesity-related measurements separately in the table 4. We tested the same association of the SNP with adiposity multiple times without correcting for these additional tests in the table 4. We recognized that this is another limitation in the study. In fact, after bonferroni correction, rs2272495 was still associated with BMI (p=0.032) and waist circumference (p=0.008). This point was added in the last paragraph of the revised discussion. We wrote, "Firstly, we tested the same association of the SNP with adiposity multiple times without correcting for these additional tests in the table 4. In fact, rs2272495 was still associated with BMI (p=0.032) and waist circumference (p=0.008) after bonferroni correction." Discussion 18. The last paragraph is unnecessary. The association with these “quantitative measures of obesity” is the result of co linearity with BMI. Response: We deleted this paragraph in the revised manuscript. Thank you. Reviewer's report Title: Genetic variations in APPL2 are associated with obesity in a Chinese population with normal glucose tolerance Version:1 Date:6 December 2011 Reviewer: Nabila Bouatia-naji Reviewer's report: JIang et al report a well designed and conducted study of APPL2 where they have tested the association of genetic variants located in this gene with obesity related traits. The rationale of the gene, variants selection and analyses are adequate. The paper is succinct and focused. The conclusions driven are appropriate as are the statistical methods. I only have minor comments: 1. Figure 1 should include more details about the gene structure to locate the variants position on exons and regulatory section. Ideally, this figure should provide all the variants and highlight the tag SNPs, so we can have an idea about the coverage. Response: Figure 1 showed linkage disequilibrium analyses for five SNPs in our study population, while we need the HapMap data to plot linkage disequilibrium for all variants. Thus, we added Figure S1 in the supplemental files (see Additional file 1). This figure shows all variants and highlights the tag SNPs selected in our study in the APPL2 locus. 2. The authors did not provide any information about the SNPs, if any, that are potentially tagged by rs2272495 (the figure with LD of all SNPs from HapMap would be useful to visualise those that are tagged and not included in the study). Response: rs2272495 can tag these SNPs: rs11831854, rs3736628, rs11112412, rs3794227, rs11112432, rs4964336, rs10861361, rs2293643, rs11112421 and rs10861359. All these SNPs were underlined in the figure S1 (see Additional file 1). 3. This is the first genetic assessment of APPL2. Reporting association with overweight and BMI and WHR is not a replication, as these traits are highly correlated. Thus these findings need further confirmation in larger studies. This should be stated in the discussion. Response: Thank you for your kind comment. Our findings need further confirmation in a larger independent sample. We added discussion on this point in the last paragraph of the revised discussion section. We wrote," Secondly, we found associations of rs2272495 and rs1107756 with overweight/obesity and obesity-related measurements without replicating our findings in a larger independent sample. Further replications in other cohorts are needed to confirm the influence of genetic variation within this locus on overweight/obesity or metabolic traits in the Chinese population." 4. The authors should provide an in silico assessment of the functionality of this coding variant. There are at least three easily available web tools to perform these assessments: PolyPhen2, SIFT and Panther. Response: rs2272495 is a non-synonymous variant which substitutes valine to alanine in exon15. This mutation is predicted to be benign with a score of 0.001 (sensitivity: 0.99; specificity: 0.14) using PolyPhen2 (http://genetics.bwh.harvard.edu/pph2/dbsnp/rs2272495.html). In the result section of the revised manuscript, we wrote, "Although this variant is predicted to be benign by PolyPhen2, whether it has effect on protein function in vivo remains unknown and should be further elucidated by mechanic studies." Accordingly, we add "Functional assessment of single nucleotide polymorphisms" in the method section. We wrote, "PolyPhen2 (Polymorphism Phenotyping v2), a revised version of PolyPhen, is a tool which can predict the possible impact of missense mutations on the structure and function of encoded proteins. The prediction is based on performing various sequence and structure analyses. We assessed the functionality of coding variant (rs2272495) using PolyPhen2. The outputs of predictions are cataloged as benign, possibly damaging or probably damaging. PolyPhen2 is available at http://genetics.bwh.harvard.edu/pph2"