The description of Gs292 needs to be greatly improved (or eliminated). The starting description gives the very scary, and incorrect, value of 5 times increased risk of Alzheimer’s for carriers of the two polymorphisms, based on a small, early study. The citation descriptions that follow don’t adequately modify this assertion. The third reference [PMID 20817350] is actually a large follow-up study by the same authors as the first study. It states that in the overall population studied the HFE and TF polymorphisms do not synergize to increase the risk of Alzheimer’s. The only exception to this is for the Northern European subset, where the synergy factor is only 1.75, not 5.1 as stated in the first reference. More importantly, in this population the HFE and TF polymorphisms only synergize to increase the risk of Alzheimer’s if the carrier also has the APOEε4 polymorphism (see Supplement Table 4). Since the incidence of APOEε4 is only ~14% (Wikipedia), most Gs292 carriers are not at any increased risk and will be mistakenly terrified, as I was. Given that this large study is the most definitive so far, I suggest that Gs292 be dropped altogether, and instead use only Gs293, which covers the consequences of the triple polymorphism.
- Thanks for alerting us to this. We agree with your main point: the initial sentence should be updated to reflect less risk for this genoset, based on more recent studies. However, the ApoE4 dependence is seen in one paper (PMID 20817350), but not seen in the other (PMID 20029940). Additionally, the "only true for Northern Europeans" may be misleading, since the only non-Northern-European population studied was the Spanish population, with far fewer cases and controls, limiting the statistical power to conclude much of anything. Lastly, the frequency of bi-carriers (carriers of at least one minor allele for both rs1049296 and rs1800562) is stated in PMID:20029940 as being 4% in AD patients, which is significantly less common than the frequency of ApoE4 carriers. More significantly, as a very quick'n'dirty calculation of expected bi-carrier frequency population-wide, ExAC shows an overall allele frequency of 0.16 for rs1049296, and 0.032 for rs1800562, leading to a possible bi-carrier frequency of 0.005, which is an order of magnitude lower than the cited frequency in AD patients. Is this conclusive? Definitely not. We also see in PMID:26827652 that the HFE-TF pairing did not pop up in a study looking for gene:gene interactions, but it's not clear if the study was statistically powerful enough to have seen it. Greg (talk) 21:48, 12 February 2016 (UTC)