Testers in Trouble

Sallis right once again.

Yes, it’s in trouble because the test stinks.

Sallis right once again – it was only a matter of time before people started waking up to the fact that the tests are only even remotely accurate only when the customer has the good fortune to be able to be matched to the appropriate parental (reference) population samples.

ALL of the commercially available ancestry testing is flawed, primarily because of the lack of sufficient parental reference samples for all populations as well as a lack of transparency with respect to interpretation of the data.

Recent experiences with “updates” to various tests supports the Sallis view, once again clearly demonstrating that I am, once again, correct.

I have been carefully monitoring sites here customers discuss their results, with an emphasis on those European-derived customers with ethnic origins in populations heretofore not well covered with reference samples.

Consider two scenarios - A and B – that have actually occurred and for which customer reports can be found. In A, the company, following SJW pressures, increases its reference population samples for non-European populations.  In B, the company following the interests of the majority of its customer base, increases to some degree (bit not enough) the scope and number of its European reference population samples.  For A, folks deriving from those unprivileged ethnies see from their updates an increase in non-European “admixture,” while in B, customers see increased European ancestry.

Amusing, in some cases, these are the same people – people tested with both companies!  Even more amusingly, some of these folks get updated for more non-European in A, while they are updated to 100% European in B.  The same exact people, with the same exact genomes. Of course, it also depends on how the testing company in question defines "European."  Labels matter as well here - in theory, results should be independent of labels, but in practice they are not.

Do you need any more evidence that Sallis is correct, and that the results of these tests are 100% absolutely dependent upon the reference populations used? And that how the companies decide to label the ancestral components is going to also determine what the stated “results” are described as?

Of course, if the tests are valid only with proper reference samples that represent the actual ancestries of the customers, then testing results (for "recent" genetic ancestry via "admixture analysis") will, ultimately, conflate with ethnic-genealogical ancestry. This being so, the latter would seem to be representative  in most cases.

The real value of genetic testing is to evaluate kinship for biopolitical purposes - not for some sort of mythical sharp dividing line (that is real only when total Identity in considered) but to compare and contrast the ultimate interests inherent in different sociopolitical alternatives.