Friday, March 2, 2018

23andMe Jumps the Shark Again

More nonsense.

Even if you think that is all good, why concentrate on countries like Nepal and Tonga, rather than concentrating on their core White customer base?  They instead should give more detailed breakdowns like North vs. South Germany, North vs. South Italy (and breaking out Sardinia and Sicily), East vs. West England, etc. The Pareto Principle folks….20% of ethnies provide 80% of your customers.
Traditionally, our Ancestry Composition report compared customers’ data to 31 different populations, allowing 23andMe to report the percentage of DNA that came from those regions. This helped connect individuals to their ancestry back 500 years before the advent of modern migration.
Poor coverage, a lack of proper interpretation, a lack of full disclosure about statistical significance, and accusations that “exotic” admixture is added to people’s profiles for leftist political agendas.

There are researchers out there with access to DNA samples from various regions, particularly in Europe, not covered in 23andMe’s “31 populations," and there are published SNP data sets.  Granted, those published SNPs are not the same ones used by 23andMe, but what about the actual samples?  What about obtaining samples yourself?  What about using verified customer data to expand the base of parental populations?  Those data already exist!  They just need to be verified, collated, and curated, and then used for the required calculations.  But no, an easy buck is made in other ways.
With this new update, we still make those estimates…
In other words, they blithely continue all their previous errors.
…but add another layer of insights with clues to the location of much more recent ancestors. 
Recent.  Now, why doesn’t 23andMe expand the parental populations for the Ancestry Composition so they can actually give their customers more accurate ancestral determinations, decrease artefactual admixture, and also decrease percentages of unassigned ancestry?

We can do this by looking for exact DNA matches between a customer and over 130,000 individuals of known ancestry from 120 regions across the globe. If a person exactly matches with five or more individuals from one of those specific regions, that region is assigned as a “recent ancestor location.”  
Five individuals!  And, how is any of this really any different from their relative finder feature?  It’s the same nonsense just repackaged differently.

If the company was interested in providing a quality product, instead of easy “buck-making” and leftist political agitation, they’d expand the number of parental populations, be more forthcoming about statistical significance, PROVIDE GENETIC KINSHIP DATA to ethnic and racial (continental) populations, allow for genetic integration analyses, and provide other breakdowns that may be of interest (e.g., for Europeans, ancestry percentages from Western Hunter-Gatherers, Neolithic Farmers, Steppe Peoples, etc.).

If I wanted to engage in Trumpian “4-D Chess” (sarcasm) I would praise 23andMe, and other currently available genetic tests, so as to prompt all the Der Movement obsessives and fetishists to critically examine those tests.  But I am unable to do this; the product is just too atrocious to even attempt to present it as positive.