Discussed at https://www.23andme.com/you/community/thread/6192/
This seems to be needing some major help, according to OpenSNP and the thread above, everybody has been genotyped either CC or no call for this SNP. In fact, in the thread 23andMe's resident scientist goes on record saying that the calls for the SNP are "not meaningful" on their platform. What exactly is meant by the wording "the minor allele is almost always in phase with the long allele" quoted both here and on 5-HTTLPR should be called into question as well - on the face of it, on a microarray it would seem to at most imply the two features are inherited from same parent. (A fact that on itself is useless as the genotyping platforms don't provide phase information).
However, later studies go even further in wording, like [PMID 21670732] "The long allele of 5HTTLPR harbours an SNP rs25531. Long alleles containing the rarer G allele are functionally equivalent to S alleles". This would certainly explain the no-calls! A hypothesis, then, is that the short form 5-HTTLPR alleles do not contain the binding sites for the SNP, and samples with one or two short copies come up as no-calls.
The same study, however, suggests an alternative for identifying long and short forms: Rs2129785 and rs11867581 with TA haplotype tag the short form. Unfortunately as there's no phase information, this can only be said for people who are homozygous for those SNP's... It would however be interesting to see how that correlates with the no-calls on rs25531. There's also no information on how to use rs25532.
There are now new studies released on this very issue, but they're behind a paywall, so maybe someone with institutional access could take a peek:
[PMID 22504458] Prediction of serotonin transporter promoter polymorphism genotypes from single nucleotide polymorphism arrays using machine learning methods
[PMID 23510579] Worldwide Population Variation and Haplotype Analysis at the Serotonin Transporter Gene SLC6A4 and Implications for Association Studies
- see File:PMID 22504458.pdf; the other study as a PDF is too large to upload so it will need to be emailed Greg (talk) 06:56, 12 April 2014 (UTC)
- Thanks. As one could've expected, [PMID 22504458] constructs a complex regression model out of 8 SNP's. The possible genotypes/genosets for those are 6561, so it's probably less useful for SNPedia, though it might be only recourse for unphased genotypes. Also the study populations being Nordic it would be good to see another study validating the model in general population before applying it widely. --Donwulff (talk) 14:52, 12 April 2014 (UTC)