Response to reviewers
Arnett et al. HyperGEN 100k genome-wide assocation and
validation
Critique: In table 2, they had better check the haplotype block in each SNP and p value in
neighboring SNP.
Response: During our preliminary analyses of the GWAS data we plotted –log(P) values
within haplotype blocks for a number of our promising SNPs. In all cases, we observed the
expected—shoulders of a peak of p values surrounding our SNP. Although we admit that
haplotype analyses may have been a viable approach in moving from the GWAS to the
validation study, we fail to see how any a posteriori haplotype analysis after GWAS SNPs
have been validated would change our conclusions.
Critique: Were any of the 12 SNPs significantly associated in the combined population (white
and African American)?
Response: We repeated the analysis with the combined race population. For log(LVMI),
SNPs rs238688 (P=0.0094), rs409045(P=0.0036), and rs4129218 (P=0.0176) were
associated. For LVH, SNPs rs409045 (P=0.0206) and rs4129218 (P=0.0209) were
associated. All of these SNPs were significant in one or the other race group in the stratified
analysis.
Critique: Results: There was 1 SNP that had different MAF in the replication sample than in
GWAS sample. How different? If it is quite different, it could indicate a genotyping error.
Response: We thank the reviewer for pointing out an error in our manuscript. The sentence
in question should have read “In the Caucasian replication sample, all but one of the SNPs
identified by GWAS (rs756529) had minor alleles that were the same as those in the GWAS
sample.” We have no evidence that there was a genotyping error.
Critique: LV hypertrophy is associated with hypertension. Thus, we wonder whether these
candidate genes were related with LV hypertrophy independent of hypertension.
Response: In the GWAS, we are unable to separate the effects of hypertension from LVH.
However, in our replications, we corrected for blood pressure suggesting that our findings
were independent of hypertension.
Critique: In table 3, ATRN and KCNB1 SNPs were only significant in African Americans.
However, these p values (0.09 or 0.04) might be not statistically significant in this population.
Thus, we cannot agree with its discussion.
Response: We agree that the language may be too strong. We have revised it accordingly.
Because African Americans were primarily recruited in a different setting, we thought they
represented a distinct cohort.
Critique: Major compulsory revision: Discussion, page 9: The authors refer to a "joint P-value",
calculated as Pgwas * Prep. This calculation is not appropriate nor is it mentioned anywhere
previously in the methods or results. It is also not clear what SNP they are referring to here. This
sentence should be removed.
Response: The sentence has been removed.
Critique: Methods: Authors note that they adjust for age, sex, weight, SBP and DM. Was height
not available?
Response: Height was corrected directly in the calculation of LV mass by indexation to
height2.7.
Critique: Were any principal component analyses conducted to determine whether there was any
significant population structure beyond self-reported race? If so, please indicate.
Response: We have conducted an analysis with the PCAs and have added a note to the
discussion. As an example of this work: for SNP rs6450415 on chromosome 5 we identified
6 principle components based on 5 million permutations. This adjustment resulted in a P
value of 6.06e-6 compared to the result in Table 2, 6.43e-6. Other SNPs showed quite
comparable results and no SNPs lost significance. We would be willing to substantially
revise the paper to add this methodology; however, this would add a substantial amount of
text with essentially no change in the results and their interpretation.
Critique: Would clarify the distribution of numbers of patients early in the Study Population
section. Would make it easier for readers.
Response: This text has been moved to earlier in the Study Population section.
Critique: Methodology – would describe the approach as a GWAS followed by a validation
study (as opposed to “two-step GWAS”. At present this portion of the paper could be written in
a more clear manner.
Response: Terminology changed throughout
Critique: Typo “populations” second last line page 8.
Response: Fixed
Critique: The Background section mentions 7 significant SNPs in stage 2 but later, in Results
there is mention of 5 significant SNPs – would clarify for consistency.
Response: Fixed
Critique: Background, par 2: The phrase "SNPs with promising significance" is vague and lacks
meaning. Do the authors mean statistically significant SNPs (as determined by an a priori
threshold) or just top SNPs?
Response: Terminology changed.
Critique: Results: The authors only present p-values in the text and in the tables. Inclusion of
betas (or ORs) and SEs are appropriate, particularly in the table to allow the reader to see the
direction and size of effect in the GWAS and replication populations. The text refers to a figure
showing effect sizes, but no such figure was in the manuscript.
Response: A figure has been added that shows the effect sizes for SNPs in the validation
where P<0.05 for LVMI. Odds ratios for significant SNPs in LVH associations are now
given in the text.