Same science, different inference

About two weeks ago I pointed to the peculiar disjunction between what a paper on Indian genetics actually said, and how people, including some of the researchers who contributed to the paper, were spinning it. For instance, the finding that South Asians can be reasonably modeled as a two-way admixture between "Ancestral North Indians" (ANI) and "Ancestral South Indians" which varies in ratio between between 7:3 and 2:3 across regions & caste groups was translated into "the genetic unity of India." And now I notice in The Guardian another Indian has an article titled Tracing the fissures in India's society: The worlds of variation discovered within the Indian genome only emphasise the difference that divides our nation:

Despite our country's much-vaunted pluralism, Indians harbour a keen sense of difference, be it of language, religion or complexion. We also often have exaggerated visions of history, or at least of myth, history's livelier twin. For instance, many Chitpavan Brahmins, a caste group in the Indian state of Maharashtra, have been known to attest their relatively fair skin to a boatload of Vikings (apparently very, very lost) who washed up centuries ago on the western coast. So it comes as little surprise to Indians that scientific research increasingly traces the roots of our diverse society to the distant past.

The latest study of the genetic history of India (detailed by Adam Rutherford) unearths worlds of variation within the Indian genome. Indians could read this new DNA evidence in a reassuring light, as confirmation of that oft-repeated cliche of India's "continental" diversity: not only do its billion-plus citizens belong to an astonishing array of linguistic and religious groups, but also India is four times more genetically diverse than Europe.

This is a position which can be supported by the findings in the paper. The same findings which some of the paper's researchers spun arguing for genetic unity. Mapping population genetics findings into plain English is more an art than a science. Something think about when people make grand assertions; e.g., "There are no human races," or, "We are all Africans." Science being what it is strong assertions are usually grounded in some genuine empirical results or analytic insight, but how one interprets those assertions may differ greatly contingent upon the database you have to work with as a filter.


More like this

On that "India is 4x more genetically diverse than Europe" -- any idea what definition of Europe is being used?

This is a simplistic question, but how much more genetically diverse is Africa than Europe? How do Europe and, say, East Asia compare. My assumption is that Europe and East Asia are probably the least diverse, while sub-Saharan Africa, South Asia, and Southeast Asia are probably the most.

This came up in a discussion where I was arguing that I, a pasty white euro-mutt American, am genetically more closely related to Obama than 95% of American blacks. My point was that most American blacks are of West African descent, while Obama was half East African -- and that there is much more genetic difference between East and West Africa than between my euro-mutt background and whatever Obama's half white background might be. (Not that I'm eager to claim kinship with a corrupt left-wing putz like Obama, but I still think I'm a closer cousin than most blacks are.)

But my friend, an educated artsy white liberal, didn't believe me, and I didn't have any specific data to cite other than the general impression I've gotten over the years. I thought this was something that was just generally known.

Is there any chart somewhere that ranks countries, regions, or official race categories by amount of genetic diversity? I'd like to see more "X is Y times more genetically diverse than Z" claims, if there is actual data to back up such claims, of course.

chris, yeah, i think you are right. in terms of total genome content you're gonna exhibit more similarity, since european:european similarity is greater than west african:east african. the main confound would be the amt. of white ancestry in a given black american. some of these charts might help:

the rank order in genetic diversity is i think:
africa > south asia & mid east > europe & southeast asia > east asia > americas