Friday, February 15, 2013

Gradient Maps for African ADMIXTURE components

Here below are gradient maps for my last African ADMIXTURE run, Africa_V2b, courtesy of a demo download of Mapviewer7 . The Kriging method was used for Gridding and 'Grid Z limits' mode was used for color mapping.

Sampled Population's Index

Sampled Population's Location

PCA for the FST distances
generated by ADMIXTURE  

West-Africa Cluster Freq.

Nilo-Saharan Cluster Freq.

East-Africa-2 Cluster Freq.

North-Africa Cluster Freq.

Khoi-San Cluster Freq.

Omotic Cluster Freq.

Mbuti-Pygmy Cluster Freq.

Biaka-Pygmy Cluster Freq.

Hadza Cluster Freq.

East-Africa-1 Cluster Freq.
Isometric view of the MDS plot
 for all Populations sampled


UPDATE (02/18/2013) : Below are gradient maps for the first African ADMIXTURE run, Africa_V1, courtesy of a demo download of Mapviewer7 . The same options as above were used both for gridding and color mapping.




Sampled Population's Index

Sampled Population's Location

PCA for the FST distances 
generated by ADMIXTURE  

Central-West-Africa Cluster Freq.

North-Africa Cluster Freq.

Eastern-Bantu  Cluster Freq.

West-Africa Cluster Freq.

East-Africa-2 Cluster Freq.

Khoi-San Cluster Freq.

Mbuti-Pygmy Cluster Freq.

Biaka-Pygmy Cluster Freq.

Hadza Cluster Freq.

East-Africa-1 Cluster Freq.

Isometric view of the MDS plot
 for all Populations sampled

22 comments:

  1. The bimodal Omotic distribution is a bit surprising. Is there any logical reason why that should be present?

    ReplyDelete
    Replies
    1. It is because of the high presence of the omotic cluster found in the Sandawe of Tanzania, possibly an ancient rift valley presence that got disrupted by incoming farmers from the west.

      Delete
  2. This test is interesting.

    I am mostly European with about 5% middle eastern in my DNA from other sites. I am according to the ethohelix with gedmatch 50% Egyptian, 25% Pakistan and 25% Moroccan.

    Now with the other DNA tests I have 20+% Italian/French region of the world so maybe it's just a carry over

    ReplyDelete
    Replies
    1. I don't understand how you could possibly get a Pakistani reference, as I have never used any Pakistanis for any of the allele frequency results generated with ADMIXTURE that I have submitted to GEDmatch.
      If you can tell me exactly which test you used to come up with those results I can give you a more helpful input.

      Delete
  3. The only Calculators on GEDmatch that include Pakistani populations are MDLP Ancient Roots K18 and Dodecad V3

    John Olson - Administrator, GEDmatch

    ReplyDelete
    Replies
    1. Thanks, that's what I thought, perhaps James Wolfe misread Palestinian for Pakistani?

      Delete
    2. Hello, Thank you for this wonderful tool! My results of DNA testing showed that I am 74% Sub Saharan west African, 1.7% East Asian & Native American and 0.2% Middle East & North African, 18% European . The EthioHelix K10 + Palestinian came up with 89.89% Palestinian. Also on the EthioHelix K10 + Japanese run 74% North Africa and 17% Japanese. The Palestinian percentage was very high. Does this mean there is a Great similarity to the West African genome?

      Delete
    3. I wish I knew how to read this data. I can't make any English out of it.

      Delete
  4. My question is do I have a connection with the Ethiopian Jews according to your findings?

    ReplyDelete
  5. Hello Ethio Helix, I was noticing something interesting with your calculator. It seems to me that when Europeans estimate their ADMIXTURE Proportion using the Ethio Helix K10 Africa Only calculator are divided into two clusters. North-Africa and East-Africa2. However, when Middle Easterners such as myself calculate our ADMIXTURE Proportions we tend to register more of the East-Africa2 and less of the North-Africa component while with Europeans this seems to have the opposite effect.

    Would this imply that Middle Easterners have a higher affinity to East African populations compared to Europeans?

    ReplyDelete
    Replies
    1. "Would this imply that Middle Easterners have a higher affinity to East African populations compared to Europeans?"

      Yes, I would tend to think that would naturally be the case. The Africa only calculator has no non-African populations used in the analysis and the fst layout of the ADMIXTURE components can be seen here. As you can see the EA2 and NA components are close to each other, but while the NA component registers highest (in terms of frequency) with the Maghrebi populations in the dataset, the EA2 component registers highest with East African populations.

      Now what happens if we start appending non-African populations to this dataset, say starting with Middle Easterners (ME), then Europeans , then South Asians and so forth, what happens is that this global genetic cline starts forming and completes itself with the addition of East Asians and Native Americans. As you can see in the cline, MEs are closer to EAs than Europeans are, therefore it is only natural that when the only available populations in the dataset are Africans, MEs would register more EA2 than Europeans.

      Now the reason to this affinity is a different ball of wax altogether, we have several lines of evidence that show after the initial OOA migrations, several bidirectional movements have occurred between EA and ME for a very long time. However, older (and even some newer) academia has taken the stance of explaining away this affinity with uni-directional punctuated movements , to be sure the direction only being from ME --> EA. I disagree with this assessment, and in my humble opinion think that it is rooted in simple old-fashioned bigotry against Africans and not reasonable logical deductions.

      Delete
    2. "Now the reason to this affinity is a different ball of wax altogether, we have several lines of evidence that show after the initial OOA migrations, several bidirectional movements have occurred between EA and ME for a very long time. However, older (and even some newer) academia has taken the stance of explaining away this affinity with uni-directional punctuated movements , to be sure the direction only being from ME --> EA. I disagree with this assessment, and in my humble opinion think that it is rooted in simple old-fashioned bigotry against Africans and not reasonable logical deductions."

      I agree, I think that the discovery of the Basal Eurasian cluster by Lazaridis et al. via the EEF component present in the Stuttgart farmer revealed undeniable African affinities. Which were later confirmed from other analyses such as Tree Mix and other formal software showing gene flow from African populations, be they Yoruba, Hadza, Dinka, etc. to Stuttgart. Which led some people to extrapolate that Basal Eurasian is either an Aboriginal Arabian component, which I personally think is the less convincing hypothesis, or that it may originally be an East African component. Which some think is far fetched, but to me this actually makes sense, since we know of an ancient harvesting community (Natufians) from the Levant being described as and clustering with African populations.

      The anthropologists at the time didn't think there was any genetic continuity between the Natufians and the present day populations in the Middle East. Now I'm starting to think they were wrong, and that the Natufians were actually rich in this Basal Eurasian component, which they contributed to Middle Easterners.

      Delete
  6. This comment has been removed by the author.

    ReplyDelete
  7. Hello Ethio Helix: what an interesting feature on gedmatch! thank you. I'm quite a layperson here, so I wish there was a 3rd grade level explanation of what I'm looking at. I'm about 2/3 European, but about 1/3 Middle Eastern. I did the French to African Proportion. A lot of small Omotic percentages; 1.5% Hadza; 2% Khoi-San; 1% Eastern Bantu; a few percentages of Nilo-Saharan, including 2.8%; and as expected some very high percentages of North African on all but 2 chromosomes. I guess my main question is, is the French aspect. Is the French % just a general percentage of Northwestern European?

    I'm not a scientist, so an easy explanation would be very much appreciated! Thank you for this very interesting feature and blog!

    ReplyDelete
    Replies
    1. ok, now I just did the Palestinian combination, and it's all over the map!

      Delete
    2. " I guess my main question is, is the French aspect. Is the French % just a general percentage of Northwestern European?"

      Yes, or European generally, the calculators are only as good as the populations referenced in the initial ADMIXTURE run, you can see those here, you can also click on the clusters to sort any specific cluster's distribution amongst the referenced populations. In the Africa + French calculator for instance, only the french are referenced amongst Europeans so they behave as a proxy for other European and close by populations, and you can even see some Africans having the French component, does that mean those Africans have actual French Admixture, No, it just means that a portion of their genome that is registered as 'French' has more similarity/Affinity to the referenced French genomes than it does to the other clusters that were found in Africa. You can think of the other clusters in the same vein.

      Delete
    3. OK, thank you for the explanation. I guess the more Southern African percentages came as a bit of a surprise for me. Very interesting! Thank you.

      Delete
  8. This comment has been removed by the author.

    ReplyDelete
  9. My results for the EthioHelix K10 Africa Only Admixture Proportions were:
    Nilo-Saharan 4.31
    East-Africa2 3.87
    Mbuti-Pygmy 1.79
    East_Africa1 1.02
    Khoi-San 2.02
    West_Africa 74.94
    Hadza 1.18
    Biaka-Pygmy 4.28
    North-Africa 4.23
    Omotic 2.36

    However, on Ancestry they were: Ghana 71%, Nigeria 13%, Benin-Togo 5%, Cameroon-Cong and Mali 2% each and Caucasian 6%.
    I would like to know if you used the Ancestry reference populations (e.g. Brong or Yoruba), and whether your website can offer more detail on the results?

    ReplyDelete
  10. So I ran the test k10+palestine and I came back 91% palestine. I'm so so so new at this and all the other tests have been saying I'm Baltic or euopean. Cam you please explain how this test is broken down in comparison.

    ReplyDelete
  11. Mine has come back the same it says 91% Palestinian. I know I have ashkanazi but I don't understand where the high percentage of the Palestinian is from. Could someone explain?

    ReplyDelete
  12. My results came back 91% palestinian too and I have Ashkenazi, perhaps it is coming from a distance ancestor ?

    ReplyDelete