Not bad, Ann, not bad. What? It only took you 3.5 years to realize the same thing I was saying back in mid-2015. A one, a one, a one-two-three….
23andMe retardation – if Jews are distinct from Europeans, then why does your company list Ashkenazi Jewish ancestry as “European?” You admit that they’re distinct yet you still place them there. This is proof positive that (1) labels given by testing companies have no connection to objective reality; and (2) they know that what they are doing is wrong, but they still do it.
At this point, I think the companies should just label ancestral components with neutral labels – e.g., “A,B, C, D, E, etc.” and then we can observe what combination of components are diagnostic of different ethnies. But, alas, while that would be more objective, it wouldn’t be good for sales now, would it?
Have fun! After all, why shouldn’t the Strzok types have access to “hater” DNA? What could go wrong?
In an email to its users on Sunday, the company’s president, Bennett Greenspan, defended the agreement with the F.B.I. but apologized for not revealing it sooner.
Another thing I’ve noticed is the following. Say that Population A is one of the categories listed on a company’s ancestral breakdowns. Population A we know is composed of individual ethnic groups U,V, X, X,Y,Z. So, A is a regional or sub-racial group. Then, the company will state that ancestry from Population C is a component of the ancestry of Population A, but then show individuals of A ancestry represented as composites of A and C.
Here we see the problem. If C is part of the genepool of A, and A is a separate category with its own designation, then how can members of A be A+C if C is already considered part of A? You can wave hands around and say that some people have more C than is the average of A, but if you follow the logic of the company’s designation that doesn’t seem like the fundamental explanation. After all, if they say that C is a component of A and is typically seen with A and if A is its own category, then any C that is part of A should simply be part of the identity of A and should not be reported separately. The C cannot be at the same time part of, and separate from, A.
I believe that the real solution to this problem is that Population A is being identified with a parental population that is mostly composed of U, or maybe U and V, and the other ethnic components of A are ignored as parental populations. If the ancestral origins of X has more C than does the origins of U, and if A is being defined as U, then X is going to equal some portion of U and some portion of C. However, the U is used to define A, so X becomes A+C, even though C is a component of A.
Solutions to this is could be to include all the components of A as parentals for A, or simply rename A as U and stop pretending that it is broader than U, or have X as a separate population label with its own parental population. But to continue pretending that although A = U it is suitable for the totality of U-Z is inappropriate, and will lead to paradoxical results and interpretations.
My question at this point is whether the companies understand this (if they do not – that is frightening) or whether they simply do not want to confuse customers (and lose potential sales by being upfront about the important nuances of interpretation - which would be contemptible).
But, hey, it’s all good for folks with names like Greenspan, so what’s not to like?