tag:blogger.com,1999:blog-7785493.post8506691737810439992..comments2024-01-04T04:11:55.717+02:00Comments on Dienekes’ Anthropology Blog: Scrubbing SardiniansDienekeshttp://www.blogger.com/profile/02082684850093948970noreply@blogger.comBlogger13125tag:blogger.com,1999:blog-7785493.post-19600441990175557692012-08-31T21:36:27.393+03:002012-08-31T21:36:27.393+03:00Without having any opinion about the African admix...Without having any opinion about the African admix among Sardinians, I want to thank you for this exercise, because I have seen so many times in genetic analyses that the globe is round and everything is relative. The principle is that we can take any two coordinates and give values for other locations using these two. If necessary we can pick third one :) . Maurihttps://www.blogger.com/profile/03670078523265515878noreply@blogger.comtag:blogger.com,1999:blog-7785493.post-39336306530285737832012-08-31T19:52:16.028+03:002012-08-31T19:52:16.028+03:00I think a file is missing to run the K3b calculato...<i>I think a file is missing to run the K3b calculator (maybe K3b.par?). It doesn't work.<br /><br /></i><br /><br />See updated link.Dienekeshttps://www.blogger.com/profile/02082684850093948970noreply@blogger.comtag:blogger.com,1999:blog-7785493.post-35144624839671843582012-08-31T18:36:02.089+03:002012-08-31T18:36:02.089+03:00I think a file is missing to run the K3b calculato...I think a file is missing to run the K3b calculator (maybe K3b.par?). It doesn't work.Acidhttps://www.blogger.com/profile/09528331665202849473noreply@blogger.comtag:blogger.com,1999:blog-7785493.post-19881827930471487472012-08-31T14:41:01.975+03:002012-08-31T14:41:01.975+03:00I won't lie that I'm an amateur to all thi...I won't lie that I'm an amateur to all this, so don't shoot me down too harshly, but, to me, the situation seems to be this: What distinguishes Sardinians from other Europeans is found here:<br /><br />http://dodecad.blogspot.co.uk/2012/04/estimating-your-gok4-related-ancestry.html<br /><br />In short, they have the lowest levels of native/pre-Neolithic European ancestry of the populations explored. I imagine a portion (but not all) of East Eurasian affinity in Europeans is proportionate to UP European ancestry: The latter were more Asian-shifted than Neolithic farmers. But does this require admixture in the formal sense? West Eurasians came later in time, as it were, than East Eurasians. West Eurasian is a subset of Eurasian -- the ancestral population -- and, because of that, distinction from that group (or from its East Eurasian cousins) is always going to be relative. <br /><br />If a population (proto-West Eurasians) splinters away from East Eurasians and then splinters again at some point, it stands to reason that, because of a multitude of factors (population size, dispersal etc.), one will differentiate from the ancestral population at a faster rate.<br /><br />If they later meet up again, one will almost certainly be more or less similar to this or that population. This is true of any group: Siberians and Americans, for instance -- if you threw them together, Siberians would be more African-shifted than Americans, despite no admixture.<br /><br />Is that not, then, what we're seeing? I don't know if it's possible to distinguish admixture from affinity, so all of what I've said may well be moot, but, to me, it's intuitive that no two related-but-separated populations would diverge from an ancestral component at precisely the same rate. <br /><br />Of course, the picture's complicated by multiple waves of later 'real' admixture, peaking in NE Europe.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-7785493.post-46678508142437669472012-08-31T14:34:14.616+03:002012-08-31T14:34:14.616+03:00This comment has been removed by the author.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-7785493.post-482851642118654712012-08-31T07:48:09.438+03:002012-08-31T07:48:09.438+03:00I think "scrubbing" is a potentially off...I think "scrubbing" is a potentially offensive adjective in this context and it would be better to choose another one. Excised, trimmed, corrected etc for example.Anonymoushttps://www.blogger.com/profile/11000684388615334278noreply@blogger.comtag:blogger.com,1999:blog-7785493.post-58035859514379401302012-08-31T00:54:32.217+03:002012-08-31T00:54:32.217+03:00The K=3 components are not assumed to be pure. The...<i>The K=3 components are not assumed to be pure. The only thing that is assumed is that if there is African ancestry in a window, it will register a non-zero African score in that window.</i><br /><br />So you <i>are</i> assuming that the West Eurasian component is devoid of African influence at the very least, no? <br /><br />What I think you have failed to bring to attention is the question about why increased East Eurasian admixture would even result in a decreased African affinity. The reason is that East Asians are further removed from Africans genetically than West Eurasians are. Americans, in turn, are even further removed from Africans. This is why an EA-mixed West Eurasian reference results in an increased African admixture estimate in populations with less East Eurasian admixture.<br /><br />It would be interesting to calculate the African admixture in the scrubbed Sardinians with an East Asian reference, to compare with the Karitiana as a reference. This could then be compared to the difference between African admixture estimates in regular Sardinians using both ancestral populations. If the difference between the estimates of both experiments is more significant when the unscrubbed Sardinians are used, then this would further support the existence of the Africa-America axis I alluded to in my previous comment (Americans are further removed from Africans than anyone else). <br /><br />Now, the compromise position in this whole debate would be that both groups (Sardinians, CEU) have some admixture that pulls them in different directions relative to a "West Eurasian" component. But simply using Northern Europeans or Sardinians as the assumed ancestral population to estimate admixture could lead to misleading results. I think the compromise position is the most viable explanation. <br /><br />I'd like to add that we still know very little about the ancestral populations who lived many thousands of years ago and how they gave rise to the variation seen in populations today. It's even possible that Sardinians may have East Eurasian influence; they simply appear to lack this in comparison to other modern West Eurasians. Not to mention the African affinity of West Eurasians in general. <br /><br />I wonder what methods the Reich lab is using in their upcoming paper.Lankhttps://www.blogger.com/profile/09164328821211694856noreply@blogger.comtag:blogger.com,1999:blog-7785493.post-35395090474270795842012-08-31T00:07:45.140+03:002012-08-31T00:07:45.140+03:00If the K=3 components are assumed to be 'pure&...<i>If the K=3 components are assumed to be 'pure', then there's not even any need for the 4-population test; simply looking at the K=3 values would suffice.</i><br /><br />The K=3 components are not assumed to be pure. The only thing that is assumed is that if there is African ancestry in a window, it will register a non-zero African score in that window.<br /><br />Given that African-ness or not of SNPs can only be assessed by comparing allele frequencies and allele correlations in different populations, I would challenge anyone to show me a segment that is of African origin that doesn't even register a 0.01% African in an ADMIXTURE test. <br /><br />What kind of scenario would result in a real African segment that consists of SNPs whose frequencies are all concordant with the frequencies in West Eurasian and Asian populations to such an extent that not even a 0.01% trace remains?<br /><br />Anyway, if anyone thinks there are African segments in the scrubbed data, I'd like to see the evidence for it.Dienekeshttps://www.blogger.com/profile/02082684850093948970noreply@blogger.comtag:blogger.com,1999:blog-7785493.post-28191033346543631092012-08-30T23:48:56.568+03:002012-08-30T23:48:56.568+03:00A ~2-fold difference in African admixture has resu...<i>A ~2-fold difference in African admixture has resulted from a different choice of outgroup. This is unexpected if West Eurasians did not exchange genes with Papuans and Karitiana since their divergence, but expected if CEU received genes from an Asian population that was more like Karitiana and less like Papuans.</i><br /><br />It's also expected if Eurasians are African-shifted, relative to Americans. You might find evidence of this African affinity in Eurasia and Melanesia, if you were to run a K=2 analysis. <br /><br />Your "scrubbing" method seems like a great method to test the effect that an Asian or African shift has on the 4-population test. However, it is limited by the use of ADMIXTURE components based on modern populations. If the K=3 components are assumed to be 'pure', then there's not even any need for the 4-population test; simply looking at the K=3 values would suffice. K=3 shows an average of 1.2% East Eurasian and 0.2% African in the CEU, and 0.9% African in the Sardinians. All Eurasians are closely related relative to Africans, which means that African admixture in Sardinians would cause a greater shift on the Asia-Africa axis than an equal amount of East Eurasian admixture in the CEU. This means that the Africa and Asia shifts of Sardinians and CEU, respectively, should roughly cancel each other out, if the 'West Eurasian' ADMIXTURE component is a proper ancestral population.<br /><br />But it would be great if you tried a K=2 run as well, to shed light on the positions of different populations on the Africa-Asia (Africa-America?) axis.Lankhttps://www.blogger.com/profile/09164328821211694856noreply@blogger.comtag:blogger.com,1999:blog-7785493.post-61307813840379192172012-08-30T22:34:45.233+03:002012-08-30T22:34:45.233+03:00Any thoughts on why Karitiana would be the better ...<i>Any thoughts on why Karitiana would be the better proxy? In another topic we already saw that Atlantic-Baltic shows some Amerindian-like component even in the absence of any Siberian-like comp.. Things are starting to get interesting.</i><br /><br />My guess is it has something to do with the Y-haplogroup R-Q common ancestry. Some old Asian group that spawned both Rs and Qs, with some of the Qs moving to the Americas ~15ka, and some of the Rs drifting towards Europe and arriving there eventually fairly recently.Dienekeshttps://www.blogger.com/profile/02082684850093948970noreply@blogger.comtag:blogger.com,1999:blog-7785493.post-88642195382567396892012-08-30T22:33:07.283+03:002012-08-30T22:33:07.283+03:00The study of Moorjani is pretty bad actually, it d...<i>The study of Moorjani is pretty bad actually, it didn't make sense all those high percentages, and is very Anglo-centered in the sense they use the CEU sample as the defacto pure European reference, whitout taking in account the northern-euro asian-like admixtures. I think your articles are pretty genius.</i><br /><br />I wouldn't go that far. It was a reasonable effort based on the available evidence. I don't think anyone could have suspected easily that east Eurasian-like admixture had taken place in northern Europe to an extent that the estimates would be affected. Also, the paper introduced a new methodology for inferring admixture times, which is still useful whether the admixture levels are perhaps a couple percent higher than they ought to be.Dienekeshttps://www.blogger.com/profile/02082684850093948970noreply@blogger.comtag:blogger.com,1999:blog-7785493.post-88955256634711457742012-08-30T21:35:02.919+03:002012-08-30T21:35:02.919+03:00The study of Moorjani is pretty bad actually, it d...The study of Moorjani is pretty bad actually, it didn't make sense all those high percentages, and is very Anglo-centered in the sense they use the CEU sample as the defacto pure European reference, whitout taking in account the northern-euro asian-like admixtures. I think your articles are pretty genius.truthhttps://www.blogger.com/profile/08622344688109770244noreply@blogger.comtag:blogger.com,1999:blog-7785493.post-48078063560276042382012-08-30T21:11:10.324+03:002012-08-30T21:11:10.324+03:00Any thoughts on why Karitiana would be the better ...Any thoughts on why Karitiana would be the better proxy? In another topic we already saw that Atlantic-Baltic shows some Amerindian-like component even in the absence of any Siberian-like comp.. Things are starting to get interesting.aramthttps://www.blogger.com/profile/04486006044411081900noreply@blogger.com