Joined: 30 Mar 2005 {Posts: 1082 } Location: New Jersey
Posted: Thu 29 Sep 2005 18:40 Post subject: U and L gradients in Iberia (was "how would interpret...")
Hi, Frank:
We've moved our office from one floor to another, hence my absence here, as I haven't had time to post at work, and I still don't have access at home. I haven't forgotten about the genetic studies...I'm starting to gather some.
I have a few questions about a study on Portugal (Pereira, 2000):
This study finds L mtDNA lineages at a rate of about 7%...about the figure you and I had discussed could be accurate for Portugal.
In one portion of the study, the following is stated:
Quote:
Although statistically significant differences were not detected between the three sub-samples considered, the geographic distribution pattern observed for U6 and L sequences strongly suggest that different population movements were responsible for their introduction into the country, although none of them had enough demographic impact to induce regional differentiation.
Does the bold portion imply that the sub-Saharan influence wasn't enough to change the phenotypical characteristics of the population? This is how I take it, and, of course, this is correct, since I can vouch for the fact that the Portuguese look like any other southern Europeans, who also have L sequences (as do some central and northern Europeans).
In another portion of the study, a graph is pictured, and underneath the chart is the following:
Quote:
Figure 4. A phylogeny of Portuguese sequences belonging to African clades L1b and L2 and to the default cluster L3* (some members of which may have a non-African origin).
What could they mean here? This L3* they are talking about...isn't it an ancestral form of the various subdivided African L3 sequences found nowadays? It seems strange to suggest a non-African origin here, although, to be fair, it does say only some L3* may not have an African origin -- nothing definite. I was under the impression that all L3's (which would include L3*) found nowadays are African, despite L3 being ancestral to all non-African populations.
The link below brings you to a page where the PDF file can be accessed, and you can look at Figure 4 yourself, and read the study, as the graph would be difficult to reproduce here. The relevant pages are 500 (Figure 4) and 501, right column, end of first paragraph.
Although statistically significant differences were not detected between the three sub-samples considered, the geographic distribution pattern observed for U6 and L sequences strongly suggest that different population movements were responsible for their introduction into the country, although none of them had enough demographic impact to induce regional differentiation.
Does the bold portion imply that the sub-Saharan influence wasn't enough to change the phenotypical characteristics of the population?
Like you, I would conclude that also. But that is not what they are saying. What they are saying is that they found no north-south gradient in L frequencies. This is important because historians agree that records show that the assimilation of former subSaharan slaves was concentrated in the south (in the Algarve region). So you would expect L sequences to be more common in the south. As I recall, the often-cited Corte-Real study, “Genetic Diversity in the Iberian Peninsula Determined from Mitochondrial Sequence Analysis,” Annals of Human Genetics, 60 (no. 4, July 1996), 331-50, found the expected north-south gradient. But then, the earlier study’s sample sizes were smaller. This highlights the importance of taking many samples.
The study’s other finding is that Berber sequences (U6) are concentrated in the north of Portugal! This just blows me (and, apparently, the authors) away. It flies in the face of all the historical records. I would love to hear an explanation of this phenomenon.
William wrote:
Quote:
Figure 4. A phylogeny of Portuguese sequences belonging to African clades L1b and L2 and to the default cluster L3* (some members of which may have a non-African origin).
What could they mean here? This L3* they are talking about...isn't it an ancestral form of the various subdivided African L3 sequences found nowadays? It seems strange to suggest a non-African origin here, although, to be fair, it does say only some L3* may not have an African origin -- nothing definite. I was under the impression that all L3's (which would include L3*) found nowadays are African, despite L3 being ancestral to all non-African populations.
I am not sure what they were trying to say. I suspect that they were saying that the ancient out-of-Africa L3* mtDNA haplotype did not become totally extinct, but that you can still find it occasionally in people whose only African connection goes back 75 kya. If you are looking for traces of the slave trade or of the Moorish occupation, a few ancient lingering L3*s do not really count.
The most important part of the study is the pairwise comparison technique that they used. It is based on the fact that random mutations in HVR-1 and HVR-2 happen at a slow but steady rate. Imagine, for instance, an uninhabited land colonized by a band who all have exactly the same mtDNA. After a few centuries, their descendants will begin to differ. A few centuries more and they will differ even more. A pairwise comparison among these people will show that the most common number of SNPs (single nucleotide differences) between any two individuals is precisely the number of mutations that occurred since the colonization event. A few (close relatives) will differ less than this, but nobody will differ more. And so, if the most common pairwise difference is, say, 4 nucleotides, and the mutation rate is one nucleotide every two centuries, then the colonization event must have happened eight centuries ago.
The starting point does not have to be initial colonization. A demographic bottleneck (like a plague or an ice age) that reduces a group to just a few members will also produce a peak in the frequency distribution of pairwise differences. Again, the number of SNP differences at the peak tells how long ago the bottleneck event was.
The best way to look at the pairwise difference graphs is to mentally flip them right-to-left (mirror image wise) and consider the slope of the graph as representing the fraction of new arrivals into a population. In this light, figure 2A shows a wave of European immigration from north of the Pyrenees four mutations ago; figure 2B shows a wave of European immigration just one or two mutations ago; figure 2C just barely detects two waves of immigrants from the Near East (two mutations and four mutations ago, respectively); 2D shows one large immigration from the Near East three mutations ago; 2E shows an abrupt but long-lasting Berber immigration starting five mutations ago; and the solid curve in 2F shows either that subSaharan immigration has been a steady flow, rather than a one-time event, or (more likely) since subSaharan Africans already had so much diversity among themselves (going back at least 200 kya), their graph is so smudged out that you cannot tell when they arrived.
Joined: 30 Mar 2005 {Posts: 1082 } Location: New Jersey
Posted: Wed 02 Nov 2005 15:12 Post subject: Portugal
Frank and I wrote:
William wrote:
Quote:
Although statistically significant differences were not detected between the three sub-samples considered, the geographic distribution pattern observed for U6 and L sequences strongly suggest that different population movements were responsible for their introduction into the country, although none of them had enough demographic impact to induce regional differentiation.
Does the bold portion imply that the sub-Saharan influence wasn't enough to change the phenotypical characteristics of the population?
[Frank]Like you, I would conclude that also. But that is not what they are saying. What they are saying is that they found no north-south gradient in L frequencies. This is important because historians agree that records show that the assimilation of former subSaharan slaves was concentrated in the south (in the Algarve region). So you would expect L sequences to be more common in the south. As I recall, the often-cited Corte-Real study, “Genetic Diversity in the Iberian Peninsula Determined from Mitochondrial Sequence Analysis,” Annals of Human Genetics, 60 (no. 4, July 1996), 331-50, found the expected north-south gradient. But then, the earlier study’s sample sizes were smaller. This highlights the importance of taking many samples.
It is strange that they didn't find any north-south L haplogroup gradient, because other studies (I'll try to find them later) found this gradient, aside from the Corte-Real one. One study found a greater frequency of L haplotypes in central Portugal, with south Portugal coming in second, and north Portugal having the least. But the majority (as I recall) show the south to have the highest (around 7-8%), followed by the center and the islands (3-5%) and then the north (around 1-2%). Every study has different samples and yields different results, so I guess all results must be taken with a grain of salt. At least one study found no L sequences at all out of a sample of 100 diverse Portuguese. Others found no Negroid Y-chromsomes, while still others found them at low levels.
What surprised me, since I thought this population was relatively isolated historically, is that sub-Saharan maternal and paternal DNA seems to be found every time Sardinians are studied. Apparently, they haven't been as isolated as one would think. Corsicans also typed with significant amounts of sub-Saharan Y-chromosomes. Sicily and mainland Italy have historically been crossroads locations, so it stands to reason that sub-Saharan admixture was found there; also, slaves were imported there by the Romans and Arabs. I guess, as you have said, the entire Mediterranean has been (and continues to be) a genetic mixing bowl.
Frank wrote:
The study’s other finding is that Berber sequences (U6) are concentrated in the north of Portugal! This just blows me (and, apparently, the authors) away. It flies in the face of all the historical records. I would love to hear an explanation of this phenomenon.
This floored me too, the first time I read it on the Racial Myths message board long ago. I immediately brought this oddity up, and no one commented on it. I guess they didn't know what to say. Strange, indeed!
Frank wrote:
I am not sure what they were trying to say. I suspect that they were saying that the ancient out-of-Africa L3* mtDNA haplotype did not become totally extinct, but that you can still find it occasionally in people whose only African connection goes back 75 kya. If you are looking for traces of the slave trade or of the Moorish occupation, a few ancient lingering L3*s do not really count.
So, L3*, according to this study, may or may not be of relatively recent sub-Saharan origin, right? This is the only time I have ever seen any indication that any L sequence may not be of recent sub-Saharan origin. I wonder if other molecular anthropologists would agree. Even if so, it still seems likely that in a population that has had historical contact with Africa (like southern Europe), and is geographically close to it, L3* would probably be indicative of rececent sub-Saharan admixture. I would guess the same would be true of cDe. In far flung, isolated populations, cDe and L3* may be residual diaspora material. What do you think?
Frank wrote:
The most important part of the study is the pairwise comparison technique that they used. It is based on the fact that random mutations in HVR-1 and HVR-2 happen at a slow but steady rate. [...]
I guess that's why those two "highly variable regions" are so good for detecting mutations and gene flow.
Frank wrote:
The starting point does not have to be initial colonization. A demographic bottleneck (like a plague or an ice age) that reduces a group to just a few members will also produce a peak in the frequency distribution of pairwise differences. Again, the number of SNP differences at the peak tells how long ago the bottleneck event was.
Didn't a bottleneck event occur in Europe? I recall reading that all European-descended folks can trace their ancestry to seven women.
Frank wrote:
[...] and the solid curve in 2F shows either that subSaharan immigration has been a steady flow, rather than a one-time event, or (more likely) since subSaharan Africans already had so much diversity among themselves (going back at least 200 kya), their graph is so smudged out that you cannot tell when they arrived.
Yes, I see what you mean. Isn't there some evidence that sub-Saharans had been mixing with Europeans consistently since the latter's arrival into Europe? As we discussed already, sickle cell measures sub-Saharan admixture in southern Europe around 10kya. I've also heard that remains of a specific ancient European tribe, the Grimaldi, had Negroid anthropological characteristics. This is interesting.
Posted: Wed 02 Nov 2005 17:56 Post subject: Re: Portugal
William wrote:
It is strange that they didn't find any north-south L haplogroup gradient...
I agree.
William wrote:
So, L3*, according to this study, may or may not be of relatively recent sub-Saharan origin, right?<snip> In far flung, isolated populations, cDe and L3* may be residual diaspora material. What do you think?
This is one of those assertions that are impossible to disprove. The low to nonexistent levels of L3 outside of Africa convince me that it went extinct outside of Africa after the diaspora. Personally, I am skeptical of their suggestion.
William wrote:
Didn't a bottleneck event occur in Europe? I recall reading that all European-descended folks can trace their ancestry to seven women.
That was in Brian Sykes's book. That Europeans descend from The Seven Daughters of Eve does not reflect a population bottleneck. The seven women existed at different times over a span of tens of thousands of years. There were many other people alive at the times that each of those women lived. They are important simply because each had a different mutation in the HVR of her mtDNA, thus founding a new traceable lineage.
It is sort of like those small villages in Wales where everyone has the same last name. It is not that the village was founded by one family nor that there was ever a bottleneck. It is simply that the number of last names in a closed population shrinks with each generation. Whenever a family has no boys their last name is lost forever. Similarly, the number of European matrilineal lineages has shrunk with each generation (some families have no girls) until today there are just seven major ones left (each with many sub-variations, of course, due to subsequent HVR mutations).
As I mentioned once before, though, there is some Eurocentrism in the classification scheme. The seven major European mtDNA haplogroups are much more similar to each other than are the hundreds of sub-groups of the Ls.
William wrote:
Isn't there some evidence that sub-Saharans had been mixing with Europeans consistently since the latter's arrival into Europe?
Yes indeed. When Europe's glaciers receded about 16kya, people came from all over to repopulate the place.
Incidentally, I am gradually adding my own file of admixture studies to your index. The next batch that I plan to put in will be the ones that discuss European admixture in Black Americans (as I promissed Salsassin).
Joined: 30 Mar 2005 {Posts: 1082 } Location: New Jersey
Posted: Wed 02 Nov 2005 19:23 Post subject: Re:
Frank wrote:
William wrote:
So, L3*, according to this study, may or may not be of relatively recent sub-Saharan origin, right?<snip> In far flung, isolated populations, cDe and L3* may be residual diaspora material. What do you think?
This is one of those assertions that are impossible to disprove. The low to nonexistent levels of L3 outside of Africa convince me that it went extinct outside of Africa after the diaspora. Personally, I am skeptical of their suggestion.
I agree with you 100%. I am skeptical, too. This is the only time in any genetic study I have ever seen any mentioning of this. I'm sure the vast majority of molecular anthropologists would agree that any L's (including all L3*'s) found in non-sub-Saharan populations are indicative of somewhat recent admixture, and not residual.
Thanks for the explanation of Sykes's book. I also agree that grouping the most diverse population on earth -- sub-Saharans -- into one "L" group is ridiculous and Eurocentric to the extreme.
Frank wrote:
Incidentally, I am gradually adding my own file of admixture studies to your index. The next batch that I plan to put in will be the ones that discuss European admixture in Black Americans (as I promissed Salsassin).
I love the way you set it up -- it looks very professional. I may be adding more later today or tomorrow, like the ones on Argentina, Uruguay, and some others on Russia, Hungary, etc. Should I post them in the Sticky as a reply, and you can add them into the index later?
I may be adding more later today or tomorrow, like the ones on Argentina, Uruguay, and some others on Russia, Hungary, etc. Should I post them in the Sticky as a reply, and you can add them into the index later?
Go ahead and add them to the main body of your original post. Also, feel free to change any of the ones I enter, if you see an error or a better way of saying something. I really like your idea of a geographical overall organization.
'
FAQ
Search
Register
Log in
