What made another GWAS for Diabetest still interesting?: "A Polymorphism Within the G6PC2 Gene Is Associated with Fasting Plasma Glucose Levels"

It was quite surprising to me to see Science published GWAS result for diabetes one after another. After reading the detail, these seems to be what make the difference.

1. In addition to the genotyped SNPs, HapMap imputation of ungenotyped SNPs seem to be on every GWAS todo list, right now -> they, however, didn't find anything new, which was not very surprising if they genotyping platform had a well-coverage of HapMap. All the tag SNPs should be there or at least were tagged with some SNPs. Doing imputation which uses info from genotyped SNPs wouldn't give much more information about untyped SNPs in this case.

2. Replication: another must do list for any good journal. Of course, unless you replicated the result. Your paper is probably going to a trash bin. Not to mention that your result have to survive genome-wide significant level (5x10-8) to be interesting. (French DESIR cohort, Northern Finland Birth Cohort,

3. They were not just looking at FPG, but also all the possible intermediate phenotypes for DM e.g. HOMA-Index for insulin secretion. It was quite upset to see no functional study result but only an extensive discussion of the gene function in glucose homeostasis, although from skimming through the text, it seems like they had a result from functional study of the gene in detail. So, I think if you are lucky enough to find a gene that already has been studied, you will have something more to discuss, and you will also can test for those intermediate phenotypes in your study population as they did in this paper.