'Blue-eyed Humans' do not 'Have A Single, Common Ancestor;

ScienceDaily has a most-retarded title up for a report on some new research, Blue-eyed Humans Have A Single, Common Ancestor. I already blogged the paper at my other blog. The paper roughly confirms the previous finding that I blogged that an SNP on the gene HERC2 might regulate expression of OCA2 so that there is depigmentation; in particular in the iris. I have another post coming up tomorrow morning on another study on HERC2 (it's in schedule).

Anyway, the title is stupid, because yes, the HERC2-OCA2 region probably has increased in frequency from a single gene copy, but blue-eyed people, and brown-eyed people, and white people and black people and Chinese people, share innumerable common ancestors within the last 10,000 years (though the number of times specific individuals show up in the genealogy varies quite a bit; we're all related, just not equally). The key point here isn't demographic, it is genetic. OCA2 is one of the longest haplotype blocks of any appreciable frequency in the European genome, suggesting massive selection within the last 10,000 years. We need to be careful about confounding conventional genealogical and demographic descriptions with evolutionary genetic dynamics driven by selection. The history of genes is not always the history of peoples, or, you're ancestors. Here a Dawkinsian distinction between replicators and vehicles is probably useful.

I know it is a pedantic point, but this sort of sloppiness makes it harder to communicate at the the intersection of genetics and other fields. Genes subject to powerful selection within the last 10,000 years, such as LCT, often sweep through huge areas and derive from one copy in one individual (as ascertained from the extraneous genetic sequence flanking the locus subject to selection). Concomitantly, there are demographic inferences one can make (e.g., a population with a higher frequency of allele X that population have a higher average fitness than another population, resulting in differential contributions to total genome content in the descendants of the two combined groups), but, over long time spans through segregation, independent assortment and recombination a particular allele will bleed out of its ancestral genetic background proportional to the power of selection. In other words, the genetic sweep can outrun the demographic wave of advance. That's why Indians from Uttar Pradesh and Swedes both carry the same variant of LCT which is probably derived from the Volga region of Russia and do not look similar at all, while Arabs carry a different variant. Despite the phylogeny of LCT, where Arabs are the outgroup, across most genes Indians are the outgroup. Over evolutionary time scales phenetic similarity does not necessarily entail genetic similarity, and, genetic similarity does not necessarily entail phenetic similarity.

Addendum: I do acknowledge that it is correct that the vast majority of of people who exhibit blue-eyes do share a common ancestor within the last 10,000 years. But I think emphasizing this point obscures the genetic facts which underly this assertion. Genomic history and genealogy are obviously related, but they are not coterminous. I think with this story it is important to get across to the typical person that one allele subject to positive selection swept through populations and increased in frequency. Alluding to the fact that the original copy was in one individual seems a trivial point; it is much more noteworthy when the same trait is underpinned by genes not identical by descent, because that defies out expectations.

Tags

More like this

Raz,

What I find strange is that the author adds:
The mutation of brown eyes to blue represents neither a positive nor a negative mutation. It is one of several mutations such as hair colour, baldness, freckles and beauty spots, which neither increases nor reduces a human's chance of survival.

As if there is no such thing as Selection...

Well, if your going to be pedantic, your title is wrong too without suitable qualification. All humans, blue-eyed or not, have a single common ancestor. It's just a question of how far back you go... ;)

Ack! If I'm going to be pedantic, I should use the correct forms of you're / your...

Need more caffeine.

Razib,

Doesn't the type of "bleeding out of it's ancestral background" necessary to blur demographics require interbreeding (gene flow)?

Just want to clarify for the "typical person".

Doesn't the type of "bleeding out of it's ancestral background" necessary to blur demographics require interbreeding (gene flow)?

yes, there is by necessity gene flow. but the power of gene flow to homogenize on a given locus (gene) is proportional to its selective value. examples

pop A <---> pop B engage in gene flow on locus 1 starting at time 0, with disjoint allele frequencies, x & y. they are of equal size.

case 1 - neutral, allele frequencies would equilibrate to 0.50 over time.

case 2 - x is selected in A, but not in B, y is selected in B, but not in A (imagine different neighboring ecologies). the different in frequency between A & B on locus 1 would be greater down the line than between the typical locus. so in this case, selection maintains difference.

case 3 - x is selected in both populations. it bleeds very rapidly from pop A to pop B, and fixes in B, replacing the allele y on locus 1 in pop B. in this case positive selection would homogenize the locus far faster than the action of equilibration through gene flow on neutral loci.

another simple illustration, you have populations a, b, c, ... etc. which are structured. there is gene flow across these populations. if there is less than 1 migrant per generation across all these demes drift can build up between group variation. but, on the loci where powerful universal selection operates the demes could remain relatively homogeneous. purifying selection would prevent local variants from coming to the fore.

Doesn't the type of "bleeding out of it's ancestral background" necessary to blur demographics require interbreeding (gene flow)?

i think i might know what you're getting at, so one more example. four individuals

1) sicilian
2) dane
3) punjabi
4) tamil

assess these individuals across loci known to exhibit variation in human populations. so now you have ancestrally informative markers. on a typical gene 1 & 2 would coalesce back to a more recent common ancestor than either would with 3 & 4. the inverse is two, generally 3 & 4 would coalesce back sooner than with 1 & 2. this is not the case with LCT, the gene which allows lactase persistence in adults. in this case, the dane and the punabji would mostly likely coalesce back to a more recent common ancestor. the tamil and sicilian would probably be different from each other and the other two groups, probably twigs on the ancestral lineages (that is, they don't necessarily have to be closely related).

so the point is that on LCT despite the fact that 80% of punjabis might share the same allele as 100% of danes, that does not mean that punjabis share 80% of the ancestors of danes. rather, selection operated at LCT so amplify the power of gene flow so that it swept across other populations and simply eliminated other alleles which were once extant.