I am finishing a PhD in biological anthropology and my research focus is the geospatial dynamics of adaptive and neutral alleles. So, I am very familiar with personal genomics technologies and data interpretation. The service 23andme provides is pretty incredible, imho, given how much phenotypic prediction we can now accomplish in such a short period of time for such low cost. A few comments:
but lactose intolerant and alcohol flush...
The giveaway is dry earwax. That SNP is essentially nonexistent outside of north Asia, though it is present at low frequencies in some southern Asian populations. Lack of any malaria resistance alleles (some of which are at considerable frequencies in south Asians) implies the OP is north Asian.
FWIW the alcohol flushing reaction is the same SNP that confers resistance to esophageal squamous cell carcinoma; interesting they present that as two distinct 'tests.'
I got a kick out of the "muscle type: Likely not a sprinter" response.
LOL!
That's a reference to
this SNP. The sprinter SNP is rare outside of sub-Saharan Africa.
Paternal testing = full genome sequencing = PCR = more man labor
23andme = SNP genotyping = SNP microarrays = less man labor
apples and oranges
Paternity testing does not require full genome sequencing. It typically requires a combination of STRs, microsatellites, and a few Mendelian markers.
The part that they are considering useless and a "scam" is the predictive part of the SNP analysis. There's nothing wrong with the raw data (what his actual SNPs are), it's just that the interpretations of GWAS/SNP analysis are still not very clear.
The data are what they are and are unlikely to be erroneous. The issue of personal genomics is that phenotypes can be both mono- and polygenic. For some things, like the dry earwax phenotype, or the ABO blood type, we know with near 100% accuracy what a person's phenotype is from their genotype. In other cases, like the Crohn's Disease phenotype, there are many loci that are associated with the phenotype, but we don't know for certain what a person's phenotype might be given specific data. Hence the variation in confidence assigned to the predictions.
I agree with the OP, in that the results can be interesting. Unfortunately, really understanding them requires at least upper-level undergraduate training in statistics and genetics. The information given by 23andme is solid, but how well can you explain anything that's cutting edge science to a layperson? The media reports on these companies, and what the companies themselves try to do, are ample evidence that there's a lot of room for the unscrupulous to exploit the ignorant in personal genomics.
Oh, btw OP, I am a homozygote for the CCR5-Δ32 variant that confers resistance to HIV. It doesn't mean I ride bareback.
😉