June 9, 2013

HERC2 haplotypes, phylogeny and frequencies

Palisto at Kurdish DNA has a most interesting report of his own production on the eye color gene HERC2, its variant haplotypes, their phylogeny and their frequency in West Eurasian and Pakistani populations.

Based on Kurdish haplotypes, he developed the following phylogeny:

All branches produce dark eye color, excepted the two colored in blue, which are associated with light eye color. 

The defining transitions from branch#3 to branch#1 are rs1129038 and rs12913832 (demonstrated to cause blue eyes in 99% of cases) while the transition to branch#2 is found at rs11636232

He also produced haplotype frequency tables for the two light eye color haplotypes (here the one sorted by branch#1 frequencies):

Branch#1 Branch#2
Brahui 2% 2%
Balochi 8% 2%
Balochi 12% 6%
Kalash 12% 16%
Sardinian 16% 4%
Palestinian 18% 3%
Burusho 18% 12%
Basque 19% 21%
Italians 25% 19%
Adygei 26% 6%
Orcadian 28% 41%
Galician 30% 17%
French 32% 30%
Russians 36% 46%
Italians 42% 27%
Swedes 42% 54%
Germans 46% 33%
Danes 52% 32%
Austrian 55% 28%
Swiss 69% 25%

In West Asia and Pakistan (the most plausible ancient origin of the trait), we see how the ancestral #1 variant is generally dominant, with the only exception of the Kalash, reaching the highest frequencies (18%) among the Burusho and Palestinians, among the studied populations. 

This pattern is continued (at overall quite higher frequencies) in Central Europe, Denmark, Italy and Galicia, with peak among the Swiss (69%). Instead the derived haplotype #2 seems dominant among Swedes, Russians and Orcadians. French and Basques are balanced for both types.

Update (Jun 25): map:

Includes also Kurdish data from Palisto's update.

The two Balochi samples are pooled in one (same weight for each), instead the two Italian samples were retained separated and assumed to be from South and North Italy respectively (not sure but makes sense). 

See also:

Update (Jun 27): Kurdish DNA just published the HERC2 data a much wider sample of populations from all Eurasia and not anymore focusing only on the blue eye haplotypes but all them instead.

It is very interesting that ht3, ancestral to blue eyes' haplotypes ht1 and, through this one, also ht2 , is widespread through the continent with very few exceptions: Russians, Belorussians, Lithuanians and a Mordvin tribe in Europe, as well as the Kurmi, Nihali, Chenchu and Puliyar in India.

Ht5 and ht6 are also very common in Eurasia, ht7 is rare in most groups but dominant in a few (Kurmi, Melanesians) while ht4 (ancestral to ht3) is rather rare as well (highest in South and Central Asia, as well as Lebanon). Other (undetermined) haplotypes are also concentrated in some populations like the Chenchu and have some importance across Asia.


  1. Palisto is wrong.

    I have AG at rs12913832 and my eyes are light blue without any brown spots.

    Probably GG at rs12913832 means slightly higher probability of blue eyes, but definitely AG at rs12913832 does not result in brown eyes.

    1. According to SNPedia: "AG at rs1291383" equals "brown eye color". I guess there may be a small chance of some other outcome but, generally speaking, Palisto does not seem to be too wrong.

      Said that, I understand that other research and particular exceptional data, like yours, must be cross-checked in order to improve understanding, but I guess that your light blue eyes must be caused by some other allele or combo.

  2. My focus is Kurdish DNA. From my observations I know that in Kurds AG at rs12913832 always results in brown eye color. This phenotype outcome is probably true for most Asian populations.
    Example: 40% of Turks (N=112) are AG or GG at rs12913832 but not 40% of Turks have blue/green eyes.

    Only in European populations with lighter skin colors AG at rs12913832 can results in green and even blue in some rare cases.

    23andme: "AG at rs12913832: IN EUROPEANS, 56% chance of brown eyes; 37% chance of green eyes; 7% chance of blue eyes."

    Thanks Maju for spreading the word.

    1. "23andme: "AG at rs12913832: IN EUROPEANS, 56% chance of brown eyes; 37% chance of green eyes; 7% chance of blue eyes.""

      Better than SNPedia. There must be some other alleles at play, what is in itself interesting.

      "Thanks Maju for spreading the word".

      No. Thanks to you for your excellent and most interesting work.

  3. Interesting little analysis here. Here are my family members results for these SNPs.


    Father: AA
    Mother: GG
    Uncle (father's brother): AG
    Me: AG
    Brother: AG


    Father: CC
    Mother: TT
    Uncle (father's brother): CT
    Me: CT
    Brother: CT

    Now my father has pure brown eyes which seems to match his results quite well. My mother has blue eyes which also seems to match her results quite well. My uncle, brother and I all have the same results for these SNPs but my uncle and I both have hazel eyes whereas my brother has pure brown eyes. That is my uncle's and my own eyes have a light brown ring around the pupil in the center and then they are green on the outside. My brother on the other hand has no green in his eyes. They are pure brown.

    1. I have the impression that hazel and green eyes imply intermediate pigmentation between brown and blue. However, unlike blue eyes, for which a single SNP variant seems to be responsible of the vast majority of cases (and almost invariably causes blue eyes), for these other color variants the exact genetic triggers are not well understood yet.

      There have been a couple of recent studies, on Cape-Verdeans and Europeans respectively, that may help in the understanding of what is known and what is still unknown about (Western) human pigmentation, including eye color:



      In the Cape Verde study, besides HERC2, there was also some notable correlation of eye color with SLC24A5, which is even more important in skin pigmentation. In the European study, a gene called SIN, appears to have also some correlation with eye color, something that was not known before. These may well be (speculatively speaking) causes of exceptions to the HERC2 "rule" and may also influence other shades of eye color.

  4. Yes I think that you are right in that the genetics behind hazel (mixed brown and green) eyes are not truly known yet. It is hard to say but do you know if hazel eyes are classified usually as being light or dark eyes? I would say dark because the brown usually stands out more and the green is only really seen in bright light conditions or up close.

  5. Maju wrote, "In West Asia and Pakistan (the most plausible ancient origin of the trait), we see how the ancestral #1 variant is generally dominant, with the only exception of the Kalash, reaching the highest frequencies (18%) among the Brahui and Palestinians, among the studied populations."

    The highest frequencies of the "Branch #1" variant among the listed populations from West Asia or Pakistan are from the Palestinians and Burushos, not the Brahuis. The Brahuis actually exhibit the lowest frequency of the "Branch #1" variant among the listed populations.

    Of course, the classification of Adygea as part of "Europe" as opposed to "West Asia" is quite arbitrary, but so be it.

    1. Indeed, my bad. A typo or lapsus happened there and I'll correct immediately. Thanks for noticing.

  6. Replies
    1. Map corrected: there was an error in the Swedish chart (h/t to Davidski).


Please, be reasonably respectful when making comments. I do not tolerate in particular sexism, racism nor homophobia. Personal attacks, manipulation and trolling are also very much unwelcome here.The author reserves the right to delete any abusive comment.

Preliminary comment moderation is... ON (your comment may take some time, maybe days or weeks to appear).