Brief reply on European genetic substructure

Rienzi seems to want to simultaneously to argue:
(1) Genetic differences between Northern and Southern Europeans are of little importance.
(2) The existence of substructure within Northern Europe is of great significance.

I'm sure Rienzi will insist that's not what he wrote. But his thought process seems clear. (Or, alternatively, Rienzi would prefer to ignore genetic differences within Europe entirely, but believes he can play "gotcha" with those who would distinguish between N. and S. Euros by pointing out variation among N. Euros; Rienzi has openly admitted as much this in the past.) Needless to say, this sort of logic doesn't work.

Rienzi acknowledges "the major genetic clustering in Europe is along the North/South axis", but doesn't seem to grasp how "major" this axis is. To put some numbers on it, in Tian et al. (2008) PC1--the "North-South" PC--explains 42.42% of the total variance in the first 10 PCs. The next largest PC takes in 8.32% of the variance, with the remaining PCs each being responsible for less than 7%.

The historic/Neolithic/Paleolithic debate is a red herring. The question is simple: how closely are various populations related today. And what do the numbers say? Southern Europeans are more similar to Ashkenazi Jews than to the Northern Europeans. From Price et al. (2008):

We computed FST statistics between clusters 1 (mostly NW), 2 (mostly SE) and 3 (mostly AJ), restricting our analysis to individuals unambiguously located in the center of each cluster (Figure 2). We obtained FST(1,2) = 0.005, FST(2,3) = 0.004 and FST(1,3) = 0.009.


No one has ever argued the above indicates Southern Europeans "are Jews"--though, in terms of genetic distance, Southern Italians and Greeks are over half-way there.

14 comments:

  1. Tian didn't look at any eastern or central Euros except Germans. No Slavs or Finno-Ugrians. You can't find the major east-west or Neolithic/Paleolithic or European/Uralic differences if you don't test the right people.

    ReplyDelete
  2. That's incorrect. Hungarians and Eastern Europeans are included in the analysis. Tian *does* find east-west differences, they just aren't "major" compared to N-S differences, or even (most likely) when compared to any of several other axes of variation in Europe.

    On the other hand, it should be noted that the "Italian" sample in the HGDP sample is from Lombardy. Thus, "Worldwide Human Relationships Inferred from Genome-Wide Patterns of Variation" tells us nothing about Middle Eastern affiliation in Southern Italians or Greeks. Tuscans show, from memory, 3% or so Middle Eastern affiliation. Some have attempted to relate this result to the ancient Etruscans, but until further evidence comes in, the two data points we have are consistent with a N-S cline of Middle Eastern ancestry in Italy.

    ReplyDelete
  3. 2 Hungarians and 6 Eastern Euros are not enough. The sample has many NW Euros and many S Euros and few E Euros, so the greatest variation isn't W-E cause the Eastern Euros add little variation.

    I will wait to see studies with representative European samples. Even in CW Europe they didn't have French, Belgians, Swiss or Austrians.

    ReplyDelete
  4. The 2008 Price study includes a sample of 60 Poles. The result is the same as in all previous studies. Poles cluster with Northern Europeans when the first two eigenvectors are examined.

    ReplyDelete
  5. from Price:

    We conclude that the top two principal components of genetic ancestry in the IBD dataset roughly correspond to a continuous cline from northwest to southeast European ancestry and an orthogonal discrete separation between Ashkenazi Jewish and southeast European ancestry (Figure 1E).

    ReplyDelete
  6. Also from Price:

    We computed FST statistics between clusters 1 (mostly NW), 2 (mostly SE) and 3 (mostly AJ), restricting our analysis to individuals unambiguously located in the center of each cluster (Figure 2). We obtained FST(1,2) = 0.005, FST(2,3) = 0.004 and FST(1,3) = 0.009.

    Keep in mind that on the first and most important axis (north-south), AJs fall along the same end of the "cline" as S. Euros.

    The fact that Spaniards and some Italians may bridge the gap between Northern and Southern clusters (as represented in two dimensions, looking at a given number of markers) does not change the FST values between Northern and Southern European populations.

    ReplyDelete
  7. I would say rather that it's Ashkenazi Jews who are half-way (or more) to being Southern European. While they're not represented in the chart below, similar Druze and Palestinians are, and both have 30% to 50% membership in the European cluster.

    http://i25.tinypic.com/28cmrlh.jpg

    ReplyDelete
  8. @8:58:

    I don't think any basis exists for saying Druze and Palestinians are autosomally similar enough to Ashkenazi to serve as stand-ins. Nor are Southern Italians or Greeks represented in the HGDP panel. We'll have to wait to see how these groups look in terms of the "Middle Eastern" cluster.

    ReplyDelete
  9. You're right. European Jews are likely to be even more genetically European than the Middle Eastern Druze and Palestinians. On the other hand, Southern Italians and Greeks are not likely to be much more Middle Eastern than Tuscans and Sardinians.

    ReplyDelete
  10. On the other hand, Southern Italians and Greeks are not likely to be much more Middle Eastern than Tuscans and Sardinians.

    You base this on? Wishful thinking?

    Neither Tuscans nor Sardinians are particularly closely related to S. Italians / Sicilians / Greeks. Like I said, as pertains specifically to S. Euro affiliation with the brown "Middle Eastern" cluster in Li, we can only wait.

    Regardless, Greeks are more closely related to their neighbors in the Near East than they are to Northern Europeans, and S. Euros are more closely related to AJs than to Northern Europeans.

    ReplyDelete
  11. All of the Italian regions in question have similar histories with regard to invaders/conquerors/settlers from the Middle East/North Africa (i.e. Etruscans, Phoenicians, Arabs). Nowhere in Italy are the markers denoting these populations (like J-M267 and E-M81) high enough to represent a significantly larger component than anywhere else. Greece is the same.

    And who's talking about Northern Europeans? I'm only addressing your (unsupported) claim that Southern Europeans are half-way to being Middle Eastern. The evidence indicates that it's the other way around.

    ReplyDelete
  12. If you had been paying attention to Italian population genetics at any point from Cavalli-Sforza forward, you would not be claiming Tuscany, Sardinia, and Southern Italy are comparable.

    My assertion was that the NE-SE genetic distance is over half the NE-AJ distance. You are unable to dispute this fact beyond blindly hoping it can be put down to European admixture in AJ.

    Regardless of whether AJ have high or low levels of European admixture, AJs are AJs and SEs are more similar to AJs than they are to NEs.

    ReplyDelete
  13. Your replies are straw men. I never claimed that those three Italian regions were genetically identical, and I never questioned the NE-SE-AJ genetic distances.

    In terms of historical Middle Eastern influences, Tuscany, Sardinia and Southern Italy show similar patterns of settlement/conquest and admixture, as I've demonstrated.

    Reporting genetic distances does not constitute an assertion. Your assertion was that Southern Europeans are "half Jewish" (read: Middle Eastern), which I proved doesn't hold water.

    Using your "reasoning," since NEs are closer to Finns than to SEs, and since Finns trace their origins to Central Asia, that makes NEs half-way to being Central Asian.

    ReplyDelete
  14. "In terms of historical Middle Eastern influences, Tuscany, Sardinia and Southern Italy show similar patterns of settlement/conquest and admixture, as I've demonstrated."

    Not sure about that.

    1) Tuscany: As for Etruscan influence, that is disputed, but Tuscans show greater membership in the "Middle Eastern" group than other European groups in the sample (which did not include Southern Italy, Greece or Anatolia). This could be due to Neolithic influence (probably greater in central than northern Italy, as the frequency of Near Eastern NRY haplogroups in Italy shows a strong north-to-south clinal increase).

    2) Southern Italy and eastern Sicily were sparsely populated during the Mesolithic, and seem to have been largely settled in prehistoric times by agriculturalists coming from Anatolia via Greece. Then there were later migrations from the Aegean, the Levant and North Africa.

    3) Like Sicily, there are likely to be geographical differences in genetic makeup in Sardinia. Middle Eastern ancestry would probably be confined to the southwest where the Phoenicians settled. So it really depends on where in Sardinia the sample(s) came from.

    I think Spencer Wells put it well- the ancient Mediterranean was a "genetic jacuzzi"- peoples from Europe and the Near East migrated and intermixed. If that bothers some folks, I'm sorry- they need to get over it.

    ReplyDelete