Yup. Just above the map there's "Assembly" then "SNP to Chrom" which says "Rev". If you hover "Rev" it clarifies that it's using the reverse orientation instead of the forward orientation. So their C and T are really G and A when on the forward orientation, which is what 23andMe uses. You're good until you assume which allele is the risk one. Only the paper can tell you that for sure. If you look at table 4 it lists both alleles, with one labeled as "coded" and the other as "other". Then in the next column it clarifies that it's looking at the "coded" allele (A) when it lists its frequency. And the beta coefficient is positive, thus A results in higher systolic blood pressure, which is what is associated with that gene (mildly) malfunctioning. Okay, you did great until assuming A is the "mutation". The gene is having its alleles reported in reverse, so the mutation is actually A to T, meaning you have the normal version. Figuring this out is quite tricky when the two alleles for the SNP are equivalent to each other when on reverse strands. So it can help to look at another missense mutation on the same gene where the alleles aren't "A and T" or "C and G". If we look at a nearby missense mutation on the same gene, such as is at location 104,595,073 (i5001477, rs104894154) we can see that on the missense line it shows G->A when it should be C->T to get the mutation. Thus we know that in a reversed SNP, the alleles are reported in reverse for the missense information, hence what looks like T to A is really A to T. Hence with AA reported by 23andMe, you have YY ([tyr]osine) version. Then if go to the variation viewer and look at the OMIM entry ( http://omim.org/entry/609300#0027 ), we can find out what's going on. The "T to A" bit doesn't help, because we don't know the orientation. But they clarify that it results in a Y to N mutation in the enzyme, hence N is the risk factor. Looking back at the original SNP page, we see that the N amino acid equates with the A allele, but the allele is reversed, hence it's the T allele from 23andMe results which result in the pathogenic N amino acid in the enzyme. There's really no indication of a problem here. You have the normal version of the 2nd SNP, and have a very common heterozygous version of the 1st SNP, which doesn't even have a big impact when it's homozygous.