Sunday, November 05, 2017

23 and Razib Khan

I blogged about my own 23 and Me analysis some months ago.

But I didn't realize how closely Razib Khan was connected to 23 and Me. He's been writing about the company often on his Gene Expression blog since 2009. He's also discussed in the Forum section of 23 and Me.

Now it's clear that Razib Khan does believe that we are all, ultimately, Africans as the site he is (or was) associated with, FTDNA Learning Center says:

Don’t we all go back to Africa?

Yes. All of our Y-chromosome lineages trace back to a common ancestor who lived in Africa at least 115,000 years ago. Some lineages migrated out of Africa; others remained.
This map shows each of the major (backbone) paternal haplogroups’ paths out of Africa.

So the real issue for Razib Khan must be: at what point did groups of people moving out of Africa become intellectually superior to those who stayed behind in Africa? What was it that prompted the increase in intelligence?

I assume it can't be as simple as moving geographically northward, since Razib Khan's ethnic ancestry is from Bengal, which is fairly far south.

If you think all this DNA info would change Khan's unshakeable belief in the reality of rigidly-defined "races" you would be wrong. The Undark piece quotes Khan:
Still, Khan insisted that his writing about the biology of race was sound. “It’s not socially acceptable to say that there might be group differences in an endophenotype — in their behavior, intelligence, anything that might have any genetic component,” Khan said. “You cannot say that, okay? If someone’s going to ask me, I’m going say, ‘It could be true.’”
Other scientists, he insisted, believe the same things. They just won’t admit it. “I’m sick of being the only fucking person that says anything,” said Khan. “I know I make people uncomfortable, but a lot of times I say what they’re thinking.”
Clearly Khan feels constrained by political correctness, not by the weakness of his Got Smarter Once Out of Africa theory.

Just this past May, Khan replied to an article Charles Murray is once again peddling junk science about race and IQ in Vox, a piece which, as much as anything, demonstrates once again what a simple-minded bigot Sam Harris is.

Murray's point, as always, is that race is everything (it's pretty much his whole career) and that anybody who denies race, and who questions the use of IQ testing, is just being politically correct.

Of course Khan defends his hero, by quoting geneticist Neil Risch admitting that they don't have any scientific definition for "race" so they just go by the US census:
Risch: Scientists always disagree! A lot of the problem is terminology. I’m not even sure what race means, people use it in many different ways. In our own studies, to avoid coming up with our own definition of race, we tend to use the definition others have employed, for example, the US census definition of race. There is also the concept of the major geographical structuring that exists in human populations—continental divisions—which has led to genetic differentiation. But if you expect absolute precision in any of these definitions, you can undermine any definitional system. Any category you come up with is going to be imperfect, but that doesn’t preclude you from using it or the fact that it has utility.
First off - how convenient that he hand-waves away the broad brush used for race identification by saying "any category is going to be imperfect." But he believes it has "utility" - but what utility could it possibly have to broadly generalize people by "race"? Unless you're already convinced that race is a real thing and so you must have something in your data results that indicates race?

Risch is later quoted as saying that there is a connection between "micro satellite markers" and which race - presumably on the US census - the subject identifies with:

He's quoted in 2005. The US census for 2000 gave these choices:

One Race - Total                              450,000
White                                         400,000
Black or African American                      10,000
American Indian and Alaska Native               5,000
Asian                                             500
Native Hawaiian and Other Pacific Islander        100
Some Other Race                                34,400

Two or more races - Total                      50,000

Thanks to genetic testing we know that a large number of African Americans are two or more "races" - something that freaks Razib Khan out as I talked about here.

The reason it freaks him out is because his own beloved 23 and Me provided data to the NYTimes demonstrating the percentage of African DNA possessed by people who self-identified as African American. The percentage is as low as 65% in some states and less than 80% in most states. In other words most "African Americans" are two or more races.

We also see a considerable portion of "whites" with one percent or more of African ancestry, with the highest  likelihood of African ancestry being in the former slave states.

Now instead of acknowledging that the micro satellite markers were unable to demonstrate the high percentage of "mix race" individuals, Khan just says this:
2005 is a long way from 2017. Risch may have changed his mind. In fact, it is probably best for him and his reputation if he has changed his mind. I wouldn’t be surprised if Risch comes out and engages in a struggle session where he disavows his copious output from 2005 and earlier defending the utilization of race as a concept in statistical genetics.
What Khan is refusing to say is that advances in genetic testing demonstrate how useless the utility of  Risch's 2005 methodology is, and instead he implies the real problem is political correctness and risk to reputation.

Tomorrow: Vox responds to Khan.