30,000 full human genomes by January 1st, 2012?

That seems to be the inference one can make from this thorough article in Nature, Human genome: Genomes by the thousand. Last summer I pointed to a projection of ~50,000 in 2011. Nature’s current tally is ~3,000 genomes sequenced. Though issues of accuracy are still important to remember, it’s pretty striking isn’t it only 10 years on from the Human Genome Project how far we’ve come? Over a one year period there will be a shift of about an order of magnitude. Though that may be a drop off from 2009-2010, which is likely to have been two orders of magnitude. Totally not rational, but I’m feeling a touch lame at only having 500,000 SNPs done.

Below are the charts from last summer….


genomeseq1
genomseq2

(link acknowledgement, Dan Vorhaus)

On the “liberal gene”

Jim Manzi has already posted on the warranted skepticism of DRD4 being reported in the press as the “liberal gene.” Here’s the original paper. The main issue I have is not with the original research, but the inevitable confusions in the media which always arise. First, we know that complex behavioral phenotypes such as religiosity and personality seem to be heritable. That is, a set of genetic variants within the population seems to track the variation in the trait (just as with I.Q.). But, it’s been a much longer haul to actually connect a specific genetic locus to said variation, though the dopamine related genes are always brought forward as candidates. Additionally, particularly when it comes to politics there’s the norm of reaction looming. One might grant that same genetic variation which predisposes Swedes in Sweden to being on the Left or the Right is operative among ethnic Swedes in Minnesota, but most of the difference is actually between population, and a function of the differing environmental milieus of the Upper Midwest and Scandinavia (though perhaps there were strong selection effects operating upon those who chose to leave Scandinavia for the USA). Finally, as with personality, there’s the problem of characterizing the phenotype in the first place in political orientation. Not insoluble in my opinion, but far less clear than something like height, or even intelligence.

The big picture is that variation on most complex behavioral traits has some upstream genetic correlates. And, we can get some sense of the magnitude (or lack thereof) of the effect in a given environment. But like fMRI the introduction of DNA probably adds more glitz than substance at this point. We’ve long known many traits which we think as purely reflective and environmental have a partial biological basis in disposition. Clearly an area to be continued….

Election 2010 Predictions

800px-SarahPalinElonFor Congress, I think that the breakdown will be:

Senate – 50 Republicans, 50 Democrats

House – 240 Republicans, 195 Democrats

My reasoning? I just took FiveThirtyEight’s numbers and shaded them a bit to the Republican side. There’s no point in making predictions unless you predict something novel and a bit off expectations. Additionally, since the readership here leans a little Left I am inclined to tweak you guys a bit and make the political Götterdämmerung even more terrifying, though I didn’t want to push my luck and give you an implausible value which you’d reject on the face of it.

Image Credit: Therealbs2002, Wikimedia Commons

Have ADMIXTURE run on your genetic data

The last 24 hours of the initial sample collection phase of the Dodecad Ancestry Project are upon us. So if you have raw 23andMe data, you got a day to send it in, if you’re of the following groups:

-Greeks (not necessarily from Greece: Cypriots, Pontic Greeks from the former USSR, North Epirotes, Griko speakers from Italy, -Muslim rumca speakers from Turkey, etc. are all accepted)
-People from the Balkans
-People from Anatolia
-People from the Caucasus
-Italians
-Non-Indo-European speakers from Europe (e.g., Finns, Hungarians, Basques)
-Scandinavians and Icelanders
-Iranians
-Armenians
-Jews from Italy, the Balkans, or Anatolia
-Assyrians
-Arabs

The point of the project is to get a better picture of genetic variation in Eurasia, especially in undersampled groups.

The global human – II

global2A reader pointed me to a second composite image of a “global human.” It is “a composite itself from four composite of Northwest European, South & West Asian, East Asian and African faces….” I was very taken aback by this face, because it was familiar: staring back at me is a younger variant of the faces of my maternal uncles! I asked a friend who has met my family their impression of the photo without a preface, and they immediately wondered if it was a stylized representation of one of my mother’s male relatives.

The possible impossibility of truth and the importance of incorrectness

In the post below on the genetic history of India, or earlier when discussing the revisions of European prehistory, one general trend that is cropping up is that the future seems more complex and muddled than we’d presumed. This introduces the real possibility that in the foreseeable future we won’t be able to opine with any credibility about the nature of the pre-literate past, because our tools are good enough to falsify simple models, but not powerful enough to distinguish between the set of more complex models. In contrast, ten years ago when it came to the expansion of farming in Europe on offer we had simple and clear dichotomies; demic diffusion of Anatolian farmers vs. cultural diffusion of farming techniques along trade routes. Ten years ago when it came to India we are mooting the possibilities between elite transmission of Indo-European language, versus demographically significant migrations into South Asia bringing the Indo-Aryan dialects.


I think that such models are wrong, because there are major parameters left out of the picture. Now in the world we see around us the possibility of really achieving plausible consensus around a positive truth has decreased significantly, because the causal possibilities are proliferating. A model then becomes synonymous with a story. But to admit that it may be that we can’t know is still a greater improvement on the delusion that we did know.

These are general observations. R. A. Fisher’s attempt to transform evolutionary biology into a deterministic set of laws as powerful as those of thermodynamics seems to have failed; at least beyond a trivial level. The importance of history and contingency, of specific detail, muddles the general insight which we can derive in evolutionary processes. But if there is no general insight to derive then we shouldn’t be deriving it, should we? False confidence in knowledge we think we have is a far greater sin than the admission of ignorance.

My Dodecad results

A few days ago Dienekes opened up the Dodecad project to a wider range of Eurasians. I decided to send my 23andMe sample to Dienekes ASAP, and the results came back today. I’m DOD075. Dienekes also just put up an explanation of the 10 ancestral components he’s generating from ADMIXTURE (along with tree-like representations of their distances). Below I’ve placed myself in the more local context of populations to which I’m close to:

ADMIXTURE10

Here are all the populations.

Karnataka is a state in northwest South India, and can be taken as somewhat representative of the Dravidian populations. The purple component is a South Asia distinctive element. Using the terminology of Reich et al. it would be the ancient stabilized hybrid population which came out of the admixture of Ancient North Indians (ANI) and Ancient South Indians (ASI). On the margins I assume there’s just noise popping out; e.g., the “East Asian” sliver among the Kannada speakers from South India. On the other hand, the Burusho have shown evidence of East Asian admixture in other studies I’ve seen. They have a bit of the derived East Asian EDAR variant for example.

As for me, no surprise that I have a lot of “East Asian” for a South Asian. Since Dienekes is more interested in Western Eurasia he didn’t go to the point of dividing the East Asians into a northern and southern branch. I’m pretty sure I’d be in the southern branch, along with the Miaozu sample (more well known as Hmong to Americans). The bigger question is how atypical for an east South Asian I am. There is a certain basal load of East Asian ancestry among northeast South Asian Indo-Aryan speakers. Another question is whether my East Asian component can be attributed to the Mundari substrate absorbed by Indo-Aryans in northeast India, or by a more recent admixture of Tibeto-Burmans. Some of both surely, but knowing my family’s long residence on the eastern margins of the Indo-Aryan speaking domains of South Asia, cheek-by-jowl with Tibeto-Burmans, I believe I am likely to have some recent Burmese ancestry. Specifically through my paternal grandfather.

Finally, though it is just as likely to be nothing, I have a bit more “Southern European” than the other South Asians. I assume this is from my great-grandfather who was from Delhi, and part of the polyglot Muslim religious intellectual class of that city. His physical type, which my maternal grandmother inherited, was clearly West Asian. He probably had non-trivial Persian or Central Asian ancestry.