Autosomal testing in the Guardian - but more questions raised than answered

I have previously posted and commented extensively on DNAPrint’s autosomal DNA race/sub-race genetic tests, and have attempted to clear up some gross misinterpretations about them.  I was therefore interested to see this article, which prompted the usual suspects to engage in one of their regular spasms of online anti-DNAP-ism.

A few comments.  The article itself is disheartening, because of a lack of complete information on the results as well as the comments of the company DNA Bioscience.  The problem is that the company is interpreting all of the minority affiliation (“admixture”) data as resulting from recent admixture; eg, grandparents, great-grandparents, “3 or 4 generations ago”, or “imported slaves”, etc.  Unless the company knows something we do not (eg, that DNAPrint has perfected the technology to “date” “admixture”, which is unlikely at the present time as it has not been announced and the costs of the tests would have presumably been increased if that was true), then their comments are embarrassingly naïve and absurd - particularly since DNAPrint makes abundantly clear on their own website that results cannot at this time distinguish recent vs. ancient events, and that even the main text of the article itself notes that the test can reveal genetic affiliations dating back thousands of years.

So why is DNA Bioscience so clueless?  Don’t they have access to DNAPrint’s ethnic data that show that individuals from, for example, India, have East Asian, Amerind, and African genetic affiliations without having East Asian great-grandparents or a family that “came to Britain via Africa.”  Again: embarrassing.  Why doesn’t DNAPrint practice some quality control over the people who they collaborate with?

With respect to the data: The African’s data was as expected, so there is not much to comment on there except for his comment that since his family has migrated all over the world (although, not presumably to Asian nations, hmmm?) that their future generations may have other DNA sequences.  In other words: intermarriage.  Contrary to the cognitively challenged who assert that miscegenation brings a reciprocal effect on genetic interests, here we once again see the reality of one way gene flow from the Third World to elsewhere.  Nigeria is not exactly a hot destination point for immigration, is it?  Thus: a net increase of sub-Saharan African EGI.

Then we have Mr. Barnett (and what is this history of that surname?) whose grandparents were Jews fleeing persecution (of course).  I find the 23% “East Asian” a bit high, in my opinion.  That’s the sort of strange result more common with ABD 2.0; but, presumably, these people were tested with ABD 2.5.  I shouldn’t be close-minded, but I suspect that the real level of East Asian genetic affiliation would be considerably less than 23%.  But the level is high enough to be very statistically significant, so some sort of Asiatic affiliation is there.  The comments by DNA Bioscience are, of course, ludicrous – the minority genetic affiliations are not from recent genealogical events.  I would like to know what kind of results East European gentiles have, particularly with respect to East Asian affiliation.  Note of course that such affiliation does not necessarily imply direct descent from East Asians (it may or may not). 

Then we have the English woman who has a borderline statistically significant “Native American” affiliation.  Contrary to comments by the company, the chances that this is from any “recent” (post-Columbus) event is remote.  Unless the company has access to certain information that is not revealed, they have NO reason to say that there is some sort of NAM ancestor in the last 150 years.  That is completely bogus.  I think the NAM result is linked to the relatively high Middle Eastern affiliation she got.  I believe that the elevated NAM that shows up in some Greeks, Italians, Jews, Pakistanis, etc. is somehow linked to some sort of Middle/Near Eastern ancestry that may be present in certain familial lines.  Is it possible then that NAM and ME “ancestry” is somehow linked?  One simple explanation would be an algorithm bug, but I suspect that it is deeper than that, and “real.”  There may be some gene frequencies shared – for whatever reason – between NAMs and Middle Easterners, probably of very ancient origin and of course predating the migration of the NAMs from Asia to the new world.  The question for DNAPrint is: is there any correlation there?  Is it that NAM present in unlikely places tends to be coinherited with ME?  Some Northern Europeans may have elevated ME – either from Neolithic
expansions or from more recent, but still ancient events (Romans in Britain?), and this ME may carry along with it a “signature” similar to NAMs.

Of interest is that Charles Kerchner tested with elevated “Middle Eastern”, and his brother had elevated NAM (Charles himself had “East Asian”; but his confidence intervals for NAM were fairly wide around “zero”).  I therefore believe that the ME- and NAM-characteristic gene frequencies, in old world Caucasians, are probably somehow linked.  How and why I do not know.  That’s DNAPrint’s job to find out.

Finally we have the South Asian who got a somewhat typical South Asian profile, except with a higher-than-expected “North European” reading.  Problem: is that 12+ % NE meaning 12% of her total ancestry, or 12% of her specifically Indo-European ancestry?  And do the error bars overlap zero?  Assuming the result is “real”; the presence of fair-complexioned people at very low levels in Asia (such as the Pakistani tribesman Sailer says looks like Eric Idle) suggests that some Paleolithic European genes may be present at a low level there.  One cannot forget the Caucasian mummies of western China, whose postulated appearance also suggests a “northern” influence.  It is possible that fair-skinned Punjabis may have a low level of Paleolithic European affiliations; furthermore, it was indicated in the article that more recent European
influence was present there.  I, however, favor the ancient explanation, the constant facile comments about recent admixture often sound strained.  I think it more likely that there are pockets in Asia that are characterized by very low levels of Paleolithic European, and slightly higher, but still low, levels of Neolithic European, influences, and in certain families these may be a bit more concentrated. 

Contrary to what some may say, the East Asian influence in South Asia has been documented in a number of studies (that I cited previously); thus that is no reason to critique the tests.  There are enough real reasons for criticism, particularly of interpretations and parental populations, without inventing more.

My advice to DNAPrint is as follows.  First, get all the middleman companies aware that not all “admixture” data result from recent events.  Not everyone who tests as “20% East Asian” has a Chinese grandmother.  That this keeps on happening is really embarrassing, and casts doubt on the intellectual bona-fides of people who really should know better.  Second, with respect to the current tests, the only way the meaning of the data can be determined is to have larger numbers of people from a variety of ethnies tested, so that patterns of genetic affiliations characteristic of various groups can be inferred.  Instead of giving away large numbers of free tests to university professors to prove “race does not exist” (a fascinating interpretation of test data that say exactly the opposite), why not strike up collaborations with population geneticists who have access to DNA samples from many different ethnic groups?  How about real scientific progress instead of politically correct nonsense? 

Furthermore, I agree with the critics in that the controversies concerning parental populations need to be resolved.  For the money people pay to have these tests done, they deserve to know some details of the process. What were the parental populations, exactly, for ABD and Euro, and how and why were they picked?  Which were picked a priori (and why) and which were picked empirically (and how).  As well, more extensive parentals can be considered for future, updated tests.

Finally, ABD and Euro need to be further refined, as suggested they may be on the DNAPrint website.  That needs to be a priority, instead of developing “Amerind 1.0”  To be blunt, what market is there for Native Americans to “discover” whether they are of Northern vs. Southern Amerind ancestry, and in what proportions?  What people really are interested in, with respect to Native American ancestry, is to determine whether they have real, direct Native American ancestry, either for genealogical purposes or to “prove” tribal membership (eg, for affirmative action advantages, casino profits, or just plain “inclusion.”).  Thus, Amerind ancestry is most lucrative to any company at the racial level (ABD); I don’t see a marked interest in subracial Amerind classification.  Thus, updating the ABD test is a priority over an Amerind subracial test.

Furthermore, since, I presume with much confidence, most present and future DNA customers are Caucasians, updating the Euro test will also be lucrative and justified, since most interest in subracial classification will come from Caucasian people.

My advice is sound from both the business as well as the scientific perspectives; however, I have little confidence that my suggestions will be followed.  More the shame that.

Posted by JW Holliday on Tuesday, September 20, 2005 at 12:16 PM in Genetics & Human Bio-Diversity
Comments (2) | Tell a friend

Comments:

Next entry: Negroes + Islam = HELL!

Previous entry: Prof. Fraser’s article now online