In this case two individuals (right) from La Braña-Arintero site in Valdelugueros (León province near the Asturian border), dated to 7000 years ago, borderline with Neolithic arrival but, according to the paper, still "Mesolithic" (i.e. Epipaleolithic).
Federico Sánchez Quinto et al., Genomic Affinities of Two 7,000-Year-Old Iberian Hunter-Gatherers. Current Biology 2012. Pay per view.
The two individuals carried haplogroup U5b2c1 (I guess that they were relatives, maybe siblings). The mtDNA sequence was retrieved in full.
It is worth mentioning that there have been other recent studies of Northern Iberian pre-Neolithic mtDNA:
- M. Lacan found mtDNA H among Epipaleolithic Basques in 2011.
- M. Hervella found mtDNA H and U5 among pre-Neolithic Cantabrians and Basques respectively a couple of months ago.
Previously, conference materials by Chandler, Sykes and Zilhao (2005) and by Kéfi (also 2005) had reported lots of likely mtDNA H and, less commonly, U5 and other lineages, in Epiplaeolithic Portugal and Late Upper Paleolithic (Oranian) Northern Morocco (Taforalt cave) respectively. Other studies in Upper Paleolithic Andalusia are also consistent with the presence of both main European grand lineages in the Iberian peninsula since Solutrean (U5 plus what?) and Magdalenian times (H with great likelihood).
Update (revised): on the nuclear DNA sequences
Interestingly the study includes a large SNP sequence of the nuclear DNA of both individuals. When compared with several modern European populations (overloaded of Finnish and Tuscans, who hijack the analysis, the result is as follows:
As we can see the dimension 1 is hijacked by the Tuscan-Finnish duality, while the dimension 2 seems almost private, extending inside all populations with minor variations. This render the PCA essentially useless and meaningless. However it would seem that the La Braña individuals cluster best with some British and some of their creole cousins from Utah (CEU).
However when global tripolar comparison is made (after removing the noisy Finns), the result is:
Interestingly the La Braña individuals not just cluster separately from Modern Europeans but they actually look "more Asian" than these. In fact, would we superimpose these PCA plots on any global one with more samples (which always takes an L shape with vortex at or near the European cluster) like this one (from Etyo Helix), the La Braña people would overlap Afghans and Gujaratis, roughly.
So it's like the analysis performed is more perplexing than informative. Not really helpful. We can only hope that the sequences are soon released to the public domain so independent researchers can perform contrasting analysis. These PCAs alone are almost useless.
Update (Jun 29): details of the La Braña-Arintero findings:
The cave was only excavated since 2006. Braña 1 was found in flexed position with ochre and a delimitation of the niche by rocks (but there is evidence that no earth was disposed on top of it). Braña 2 was found in non-anatomical disposal (possibly because of post burial alterations by water or people), at the bottom of a natural hole 4 m. deep, just beside the other one and with indications of being also a primary burial.
Braña 1 had no caries but erosion of some teeth in what is suggested to have been caused by the use of teeth tool. There is a serious injury in the upper jaw by piercing artifact (spear?), which healed.
Associated to Braña 2, there were found 24 perforated deer canines, possibly part of a necklace or other decoration originally.
The two burials are from almost exactly 7000 years ago. The ages are:
Refs.
Julio Vidal et al., LOS HOMBRES MESOLÍTICOS DE LA BRAÑA-ARINTERO(VALDELUGUEROS, LEÓN): UN HALLAZGO FUNERARIO EXCEPCIONALEN LA VERTIENTE MERIDIONAL DE LA CORDILLERA CANTÁBRICA. 2008. ··> found at Academia.edu.
Julio Vidal et al., LOS HOMBRES MESOLÍTICOS DE LA BRAÑA-ARINTERO (VALDELUGUEROS, LEÓN): EL HALLAZGO, SITUACIÓN, ASPECTOS ARQUEO-ANTROPOLÓGICOS, CRONOLOGÍA Y CONTEXTO CULTURAL ··> found at Academia.edu.
 |
| Pre-Neolithic mtDNA lineages found in Northern Iberia to date (all quite unquestionable) |
Update (revised): on the nuclear DNA sequences
Interestingly the study includes a large SNP sequence of the nuclear DNA of both individuals. When compared with several modern European populations (overloaded of Finnish and Tuscans, who hijack the analysis, the result is as follows:
 |
| Figure S3 PCA with 1KGPomni Chip Data (European Populations) and La Braña 1 (left) and La Braña 2 (right) click to expand to original resolution (still not enough to discern Iberians by subpopulation easily) |
As we can see the dimension 1 is hijacked by the Tuscan-Finnish duality, while the dimension 2 seems almost private, extending inside all populations with minor variations. This render the PCA essentially useless and meaningless. However it would seem that the La Braña individuals cluster best with some British and some of their creole cousins from Utah (CEU).
However when global tripolar comparison is made (after removing the noisy Finns), the result is:
 |
| Figure S2. PCA with the Shotgun Data from La Braña 1 (Left) and La Braña 2 (Right) and the Worldwide Data Set from 1,000 Genomes Project, |
Interestingly the La Braña individuals not just cluster separately from Modern Europeans but they actually look "more Asian" than these. In fact, would we superimpose these PCA plots on any global one with more samples (which always takes an L shape with vortex at or near the European cluster) like this one (from Etyo Helix), the La Braña people would overlap Afghans and Gujaratis, roughly.
So it's like the analysis performed is more perplexing than informative. Not really helpful. We can only hope that the sequences are soon released to the public domain so independent researchers can perform contrasting analysis. These PCAs alone are almost useless.
Update (Jun 29): details of the La Braña-Arintero findings:
 |
| Close-up of Braña 1 (source) |
Braña 1 had no caries but erosion of some teeth in what is suggested to have been caused by the use of teeth tool. There is a serious injury in the upper jaw by piercing artifact (spear?), which healed.
Associated to Braña 2, there were found 24 perforated deer canines, possibly part of a necklace or other decoration originally.
The two burials are from almost exactly 7000 years ago. The ages are:
- Braña 1: 6980±50 calBP (5990-5740 BP - raw C14)
- Braña 2: 7030±50 calBP (6010-5800 BP - raw C14)
- El Espertín (Burón): 7080±40 BP and 7790±120 BP (not calibrated)
- Los Canes: unspecified dates from the 6th millennium BP (not calibrated), where similar deer canine decorations were found also in burial context
Refs.
Julio Vidal et al., LOS HOMBRES MESOLÍTICOS DE LA BRAÑA-ARINTERO(VALDELUGUEROS, LEÓN): UN HALLAZGO FUNERARIO EXCEPCIONALEN LA VERTIENTE MERIDIONAL DE LA CORDILLERA CANTÁBRICA. 2008. ··> found at Academia.edu.
Julio Vidal et al., LOS HOMBRES MESOLÍTICOS DE LA BRAÑA-ARINTERO (VALDELUGUEROS, LEÓN): EL HALLAZGO, SITUACIÓN, ASPECTOS ARQUEO-ANTROPOLÓGICOS, CRONOLOGÍA Y CONTEXTO CULTURAL ··> found at Academia.edu.
Update (Jul 1):
Even if I do not often see eye-to-eye with Dienekes, he is the one providing the first known independent analysis of the La Braña people. He has done so in two successive updates of the entry on these interesting, yet perplexing, sequences.
As you may know, he uses two main sets of zombie components (deduced from his own analysis, sometimes a bit questionably but well...), which he calls K7 and K12. They are Eurocentric in concept, what is convenient for this purpose.
K7 has three West Eurasian components that he labels West Asian, Southern and Atlantic-Baltic. Where Atlantic-Baltic is the main European component, being strongest towards the West and the North, the top population being Lithuanians and the 50% isocline going through the Tyrrhenian Sea, the Northern Appenines, the Greek-Bulgarian border and the ethnic border between Russians and others in the North Caucasus. In other words: Europeans are generally high in this component, with very few exceptions, all them Mediterranean or Caucasian.
For the K7 analysis, Dienekes got:
9.3% African and 90.7% Atlantic_Baltic
No modern Europeans are so high in the Atlantic-Baltic component (Lithuanians are 84-88%) but more notably none have such high levels of (simplified) African component, the closest being modern Portuguese with 2.5%. In fact Portuguese-D are in Dienekes' K7 analysis: 60% Atlantic-Baltic, 2.5% African and the rest belonging to components that we can safely consider Neolithic or post-Neolithic arrivals:
- 30% Southern (i.e. Eastern Mediterranean South or Arabia)
- 7% West Asian (i.e. Eastern Mediterranean North or Caucasus)
- 1% South Asian
After removal of these likely Neolithic-arrival components and apportioning, the "reformed" Portuguese sample would end up being 96% Atlantic-Baltic and 4% African, what leads me to think that some extra Atlantic-Baltic may also have arrived with the Neolithic-specific components, assuming that the La Braña pre-Neolithic substrate can be extrapolated to Portugal (what may be but is not necessarily the case).
As for K12, it may shed even more light. According to Dienekes, the La Braña people produce:
45% Atlantic_Med, 41.6% North_European, 10.3% East_African, 1% Sub_Saharan
Basically the Atlantic-Baltic has split in two European subcomponents and the African one in two as well (with a tiny 3 points displacement from the European to the African components as these become more refined). Remember that they are zombies (pre-determined components) and not the product of any direct comparison with actual populations. However for this purpose it is probably the best thing to do initially.
For reference the Atlantic-Mediterranean component is highest in SW Europe (Basques and Sardinians notably), while the North European is highest in NE Europe, in spite of the misleading name (former USSR, Finland). British (or the CEU HapMap sample or the Irish) are almost half-and-half for both components, just like the La Braña people, what explains why they cluster best with them... if we ignore the African part.
Like the La Braña people, British also have c. 10% "other stuff", in their case it is however the "Gedrosia" component, strongest in Pakistan and Caucasus and which is suspected to be informative of Indoeuropean (Kurgan) ancestry (along with uncertain amounts of the Northeastern, "North European", component).
But the big mystery is the East African component, which is distinct from the North West African (Moorish or Berber) one, which we might all have expected on first look. This East African component is highest among the Sandawe (68%), Somali (69%), Ethiopians (55%) and that's about it. North Africans have some but not at all at such levels and certainly not without the NW African or other components absent in the La Braña pair.
This is extremely hard to explain because modern Portuguese's African component is nearly all NW African (7.7%), with some minor Sub-Saharan (0.7%) and nearly no East African (0.1%). So the illusion maybe generated at K7 of these people's genomes being explainable by means of revising modern Iberians... seems to collapse.
So the results remain perplexing (I have received criticisms for using this and similar words but that's the truth and if you look online, nobody seems to have good answers - it's not just me).
But the results are still informative. Regardless of whatever the East African component means (what obviously demands further analysis), it seems obvious now that both European-specific components (North European and Atlantic-Med in Dienekes' slang) are pre-Neolithic.
And that is important because it opposes the most common and simplistic Neolithic replacement models. It would seem that, whatever the intra-European genetic shaking and re-scattering and the arrival of minority genetic components from the Neolithic onwards (mostly from West Asia), the bulk of European ancestry is Paleolithic European after all.
Update (Jul 1 - II):
 |
| (From Dienekes' blog - fair use of copyrighted work) |
Dienekes is even providing us with a nice visual comparison (right), using his K12 zombies, of five prehistoric Europeans.
Notice that the two so-called hunter-gatherers from Götland (an island off Sweden) are actually Neolithic hunter-gatherers at the best and Neolithic mixed-economy people most likely. They belong to the Pitted Ware culture, which is derived from the Neolithic of Ukraine and nearby parts of Russia: the Dniepr-Don culture. Derived from this, eventually some people with characteristic pottery, possibly some domestic animals such as pigs, but a largely gatherer economy (hunt and fish) spread towards the Baltic and in some cases crossed it towards Sweden.
That would explain their high levels of NE European (aka North European) component.
The South (or rather SW) Swedish farmer belonged instead to the Megalithic phenomenon and the Funnelbeaker (TRB) culture of Denmark and nearby areas. His affinities are mostly Sardo-Basque (Atlantic-Med), with a bit of Neolithic and NE European elements.
The North Italian farmer (or rather Alpine pastoralist) is the famous Ötzi "the iceman". His affinities are also Sardo-Basque (to a lesser extent than the Nordic farmer) and also has a good deal of West Asian affinities (Caucasus and SW Asian) and even some North African affinities.
Finally the Northwes Spanish hunter-gatherer is a combo of both La Braña genomes (very similar and quite incomplete), whose meaning (in my opinion always) I have discussed in the previous update section, a few lines above.
Updates (Jul 3): strange ritual for 'Mesolithic' and some other DNA comparisons
The latest zombie comparison (different set of zombies) by Dienekes (scroll for updates in the already provided link) is similar to the previous (i.e. c. 90% like modern Europeans, specifically NW ones without the "Neolithic" stuff) but instead of being anomalous towards Africa it is towards East Asia. Not sure what to think other than it seems to point to some unchartered component with deep roots.
Maybe more interesting but still in the hmmmm zone is the comment at Diario de León[es] by Pablo Arias, Director of the Prehistoric Research Institute of Cantabria. Among other less important details,he argues that the burial or rather lack of it thereof (as the corpses were placed in the cave without earth covering, even if Braña 1 does display signs of intentional funerary disposition) is most anomalous, actually unique, among 'Mesolithic' burial practices which always included proper burial and not just disposition. My translation:
And... what we just needed, so to say: Bushman affinities! In his latest comparison, using the DIY Harappa World calculator, Dienekes finds that the Braña composite is stubbornly >80% European (in this case 55% NE Euro and 27% Mediterranean) and some 17% (the highest figure so far!) non-European, mostly 7% Siberian and 6% Bushman (San).
Updates (Jul 3): strange ritual for 'Mesolithic' and some other DNA comparisons
The latest zombie comparison (different set of zombies) by Dienekes (scroll for updates in the already provided link) is similar to the previous (i.e. c. 90% like modern Europeans, specifically NW ones without the "Neolithic" stuff) but instead of being anomalous towards Africa it is towards East Asia. Not sure what to think other than it seems to point to some unchartered component with deep roots.
Maybe more interesting but still in the hmmmm zone is the comment at Diario de León[es] by Pablo Arias, Director of the Prehistoric Research Institute of Cantabria. Among other less important details,he argues that the burial or rather lack of it thereof (as the corpses were placed in the cave without earth covering, even if Braña 1 does display signs of intentional funerary disposition) is most anomalous, actually unique, among 'Mesolithic' burial practices which always included proper burial and not just disposition. My translation:
Nevertheless, he details that there are some peculiarities making it unique: like the kind of necklace found along individual number 2 and very specially the characteristic of the funerary space: corpses deposited on the floor of a selected hypogeum. "From another point of view, La Braña-Arintero provides another evidence of the spectacular increase in the number of burials in the 7th millennium a.C. and that may relfect an intensification of the territoriality in these societies".
Pablo Arias precises that the more striking characteristic of the funerary context of La Braña-Arintero resides in the sepulchral space itself, detailing that it is an exclusively funerary site, with no link to settlements of that age. We see that a remote cavity, apparently not suited for habitation, and the corpses have been placed there in peculiar spaces, well delimited by natural space in the way of niches.
And this, as he defends in the article, is a funerary behavior that has no clear precedents in other peninsular contexts and rather reminds to behaviors more common in later periods instead.
And... what we just needed, so to say: Bushman affinities! In his latest comparison, using the DIY Harappa World calculator, Dienekes finds that the Braña composite is stubbornly >80% European (in this case 55% NE Euro and 27% Mediterranean) and some 17% (the highest figure so far!) non-European, mostly 7% Siberian and 6% Bushman (San).
Haven't they compared the mesolithics with Basques? It seemed to me they did. That's what I've read in Dienekes' blog:
ReplyDelete"In the latter PCA, where the origin of each Iberian sample is known, it is possible to see that the Mesolithic specimens are not related to modern Basques, contrary to what has been previously suggested in some recent studies"
Yet I don't understand why they say these two speciments are more related to northern than to southern Europeans. Finns don't seem specially related to them.
I don't see Basques being tested either ...
ReplyDeleteI don't get why Dienekes concludes modern-day Iberians (well, it's just about one ancient Asturo-Leonese sample) are not the descendants of Mesolithic Iberians : actually, the results are pretty consistent with what we know of the subsequent migrations affecting the Iberian peninsula (Med/West Asian migrations).
I'll admit there's indeed a problem with the Admixture tests Dienekes is running with all those ancient people constantly appearing to be nearly fully "Atlantic/Baltic" as he labels the cluster bar that one individual from South Sweden which show a balance between the main two European components (quite similar to modern Basques in his tests) : Atlantic/Baltic and Med.
Should that mean that indeed, in early Neolithic times, up to modern-day Basque lands were subject to Med migrations preceding West Asian ones (read : Anatolia then Balkans) ? I find it strange.
@Neandertalerin,
ReplyDeleteThe 2 specimens are more related to Northern Europeans, as they are closest to the British, who are Northern European??
What's not clear about that?
@pconroy:
ReplyDeleteIn those graphs, I only see 3 northern european populations (two of them very closely related to each other) and 2 southern ones.
Finns don't seem closer to the mesolithic individuals than Tuscans are. The closest are CEU/British, but northern europeans also include Germans, Latvians, Russians, Danish, etc. What's even more: CEU/Brits are closer to southern europeans than to the Finns, so I don't understand where does the division between northern and southern europeans come from.
When I think of Northern Europeans, I think of Irish, British, Scandinavians and Lithuanians. I exclude the Finns and Estonians are they are North East Europeans and quite different. The Slavs are more Eastern European then Northern.
ReplyDeleteMuch of Germany is Central Europe and is closer to Northern Italy than Northern Europe.
I don't know how you guy can read that PCA plot, the vertical amplitude of Britons is too considerable to properly analyze something. It is not readable at all. Only Tuscans seem differentiated. The Mesolithic Iberian appears to be neutral on both axes.
ReplyDelete@maju
ReplyDeleteI heard its in the "mini library".
I have no idea what that means though.
If you heard that from Jean Manco, she means she has it in her "cloud" Google library, which she shares with a number of people.
DeleteYes.
DeleteYou knew where I got that from. Do you visit world families?
No but from the words used I deduced... If you ask her she will surely agree to allow you access, I guess. It's nothing but her personal "cloud" collection of papers. I always forget to organize mine... :)
DeleteLots of comments in so little time! :)
ReplyDeleteI must say that, on second thought, the PCA is very much hijacked by specially the Finnish-Mediterranean duality (PC1) and the La Braña samples fall near the zero, what is not what I would expect from someone very different from modern Europeans as a whole, but rather someone quite average, even if not similar to the local populations maybe.
But then in the continental comparison plot they fall apart, tending towards "India" (or where I'd expect to see West Asia, India and Central Asia in similar better sampled comparisons).
So in the European only comparison, in spite of the Tuscan-Finnish hijacking of the plot, we can conclude that the individuals look quite average European, out of the Asian-influenced extremes that these two populations represent (influenced by West Asia in the TSI case and Siberia in the Finnish case).
However in the global plot we are pushed to think that these people are somewhat "more Asian" than modern Europeans.
I have got my hands on a copy and there's not much more that I can see other than in the Supp. Material there's a comparison with a more varied Iberian sample, including some Basques. The results are very similar and it's extremely difficult to discern which Iberian subsample is each because they are plotted all in yellow, which almost does not contrast against the white background.
If you want a copy just email me: lialdamiz[at]gmail[dot]com
Edited the nDNA part (the update) and bumped the entry (it may appear twice therefore in your RSS feeds).
ReplyDeleteVery interesting Maju, In the past they also found E-V13 in Iberia from 7,000 years ago, source: "Ancient DNA suggests the leading role played by men in the Neolithic dissemination", presuming this is around the same time frame, they had proposed the E-V13 as a 'Neolithic' lineage, yet the PCA on the autosomal DNA of these new specimen from about the same time frame does not look very close to where the 'traditional' place of where the Neolithic dispersion (at least one of them or the main one) is thought to have originated. Does this mean that this 7KYA is the transition time for Iberia? Can you shed some light on this?
ReplyDeleteIndeed, according to the latest calibrated dates, c. 7500-6500 BP was the Neolithic transition in SW Europe (including much of France AFAIK). See this entry.
DeleteThe oldest dates in Iberia and France for Neolithic (variants of Cardium Pottery) are all from around 7500, always in the coastal areas (excepted some anomalous dates from Aragon which are probably wrong in something - they appear to be slightly older than anything else).
These individuals are certainly borderline, per the chronology, but I have no reason so far to think they would be Neolithic instead of Epipaleolithic.
I can't see the diagram at all it is very small, do you have a bigger picture? thanks
DeleteNo, sorry: I did all I could with the original image, namely marking the thousands with red bars.
DeleteThe explanations (and another chronology table, and link to the original PDF in Spanish) are in the original entry, which is the link I actually meant to post in the previous comment (my apologies).
Is it possible that these two mesolithics are actually mixed with neolithics or are neolithics themselves? That'd explain why they're more East Asian than present day Europeans.
ReplyDeleteIf these mesolithics were descended from the first HAM that arrived in Europe about 40 Ka, I wouldn't expect them to be closer to E. Asians than the middle easterners who arrived there much later. Alternatively, maybe there were multiple waves of immigrants from Asia to Europe in the last 40.000 years, not just two or three.
I agree with your perplexity, Maria Llüisa. I have no good answers but naturally I am rather distrustful of the usefulness of the comparisons performed by the research team.
DeleteThe individuals are borderline with Neolithic but I have no specific reason to doubt that they belonged to a purely Epipaleolithic community.
Very borderline in fact. I have just updated (again) with some further data of the findings and, while it looks somewhat reasonable that these are 'Mesolithic' findings, the evidence is at least a bit inconclusive, as nearly the only reason is the chronology, which the authors say is "unquestionably" 'Mesolithic' but is actually "unquestionably" borderline.
DeleteIt looks like the samples are closest to a hypothetical people that are a hybrid of British and finns from the PCA plots, PC1 looks like an North East to South/Western axis, separating the finns from other Euros, PC2 looks like a North West to South East Axis, separating British from Tuscan, so on PC1 this 7 KYA sample looks almost center but leaning more to the finns, while on PC2 looks leaning more to the British than the Tuscans.
ReplyDeleteThe closest individuals (in the European PCA) are some Utahns (CEU, blue triangles) in fact, but then in the global plot no modern European is close at all.
DeleteSomething most people don't seem to understand is that zero means nothing in a PCA, unless it is demonstrated by other data: zero is neutral as to the extremes of the selected PCs, so the La Braña people were neutral to PC1 (neither high nor low in the "Mediterranean vs. Finnish" axis, although slightly trending to the negative side, as we would expect from any pre-Neolithic peoples) and neutral also to PC2 - whatever it represents, to me it means nothing, specially with La Braña 2 - only in the left plot, La Braña 1, actually behaves as you suggest: Britain-TSI (what may again suggest slight non-Mediterraneanness, what is logical).
In this case, I'd create zombies and run supervised Admixture on the samples in order to detect real affinities. Here we only see neutrality to the displayed European trends (PCs).
In a different example (for comparison): Indians, when in a global PCA, behave as non-African but relatively neutral in the West-East axis (tending to Europe but not too much). That's not because they are admixed or anything (they tend to cluster with truly admixed peoples like Uyghurs but no relation in fact) but because neither axis captures their "essence", so to say.
DeleteThis one only shows up in Eurasian-only PCA plots, when they define a whole axis, against which both European and Chinese are negative.
That's why PCAs are often misleading... they can be useful but only when properly designed. And these are not. And the only thing this paper offers is PCAs...
Ah, I did not see the global plots, in that they would maybe be close to modern Chuvash type people?
DeleteI agree that an intermediate appearance in a PCA plot does not necessarily indicate a mixing event between distinct populations A and B, it could just as well have been due to descent from a population that had common ancestry between populations A and B for instance, that is why I used 'hypothetical'
I had not seen the Chuvash over there but more like the nearby (in the plot) Sindhi than the Chuvash.
DeleteWhy? because in the PC1 they look "less Eurasian" (or "more African" if you wish) than any European or East Asian sample, a (false, illusory) tendency that does happen with Indian samples but not with truly admixed West-East Eurasian ones like Chuvash and the like.
It's a tendency that it's just saying: I don't really fit well in this comparison: the reference points are confusing. I'm not saying that the La Braña people have actually any connection with Sindhis or anything, just that they display the same kind of neutralization tendencies in a simplified tripartite global polarity.
More detailed comparisons are needed before we can say something more clear.
Yes, there is a slight African pull on PC1, I think your assessment is reasonable without having enough data, the first sample for instance is approximately 1/5 of the way from the center of Europeans to the center of East Asians on the Y Axis, while it is about 1/20 of the way between the center of Europeans and the center of the Africans (Luhya and Yoruba in this case) on the X axis, offsetting with these approximate coordinates I see on my PCA plot populations such as Makrani, Balochi, Brahui, Kalash and Pathan.
DeleteIndeed. But, IF the Braña people would truly have a West Asian or Indian tendency (as would seem from this simplistic comparison of global plots) then they would show up as more "Tuscan" in the European plot, because generally Mediterranean peoples with some apparent West Asian admixture (as is the case of Tuscans) tend to also have more of the "Pakistani" components (Balochi or whatever you call them).
DeleteThat does not happen at all.
It's weird: the PCAs are more confusing than clarifying.
From an older post: "That alone should have put the Neolithic replacement theory to rest but, of course, faith is blind and irrational, and some seem to be grasping for the proverbial straws..."
ReplyDeleteIt's funny to observe how the same is now happening to the proponents of the Paleolithic Continuity theory. Why not just acknowledge the new evidence against your theory, and stop calling the paper confusing, perplexing, and unhelpful?
I call the paper, or more precisely the "analysis" performed, whatever I deem appropriate.
DeleteWhile the authors extracted nDNA from borderline Epipaleolithic-Neolithic individuals (the data is not as clear as we'd like in this matter) the main issue is how they compared with modern populations and that is quite sub-par.
I await independent reanalysis and hope it is done as soon as possible.
Whatever the case you can't say as Lalueza has been saying all around that this paper (without challenging methods and conclusions) demonstrates that the cave people of La Braña are not ancestors of modern Iberians, IF anything it'd suggest that:
1. Modern Iberians would be less directly related, maybe mixed with some other peoples.
2. Modern Brits (and white Utahns) are their descendants (well, not really but one could easily construct that narrative, specially considering that there is some archaological evidence for Neolithic demic replacement in Britain (and later in Utah).
Even accepting the narrative construed in the paper, which I find oddly confusing, there's a lot that does not make good sense and require further exploration. Only a pseudoscientist (someone who considers science as means to an end, instead of as the end itself) would argue against that.
I don't think that this paper demonstrates much in either direction. On the other hand repeated finding of pre-Neolithic mtDNA H (in Iberia of all places!), that really challenges at least the most common narratives of the Neolithic Replacement hypothesis.
Maybe if you can construct a model B...
But I would not expect so much from someone who signs as "Unknown" (i.e. someone we all probably know).
"Interestingly the La Braña individuals not just cluster separately from Modern Europeans but they actually look "more Asian" than these."
ReplyDeleteActually, dienekes was wrong about that. Brana is closer to Asians AND Africans. So... I guess that means the Europeans have been drifting further from the two.
Maju, I got good news for you. Look at this: http://www.worldfamilies.net/forum/index.php?topic=10663.msg133689#msg133689
"In terms of closeness it goes roughly NW Europeans>Western Scandinavians/Germans>Iberians>North Italians>Eastern Europeans>NorthEastern Europeans>Greeks and South Italians."
-jeanL
Since brana shares more with the west then maybe western meso lineages did survive.
I do not have a fully finished opinion, unlike so many others, relative to which news can be "good" or "bad"... but I would agree that the Dienekes' analysis (scroll down to updates) is interesting.
DeleteAccording to that, in the K7 analysis, they are: "9.3% African and 90.7% Atlantic_Baltic". Atlantic-Baltic being a component that is very high among North Europeans but also among Basques (73%).
And in the K12 analysis, they are: "45% Atlantic_Med, 41.6% North_European, 10.3% East_African, 1% Sub_Saharan". What, per the relevant ref. spreadsheet. The Atlantic-Mediterranean component is highest among Sardinians and Basques, clearly indicating that, if these ancient people are Epipaleolithic the AM component is as well (it was argued it being Neolithic). It is a widespread component through Western Europe (40-60% levels).
The North European component is most dense today among Lithuanians in fact and rather than "Northerner" it is a Northeasterner European component, reaching levels above 50% (actually above 60% and even 70%) in NE Europe only (former European USSR, except Caucasus, and Finland). For reference the British are at 40-50% levels, while the French and Spanish are under 40% and under 30% respectively (Basques are under 20%, lower than Spaniards).
The most intriguing feature is the "NE African" and "African" components, which I feel legitimate to suspect as of North African origin, very specially because the people of NW Iberia do have notable NW African (and other less clear African) Y-DNA and mtDNA lineages, whose origin in the Solutrean or Neolithic I could not clearly discern before now.
Ref. K7 and K12 "zombies" according to Dienekes.
The sequences are here. Now I wish I knew how to manipulate that stuff so it can be freely compared with a varied array of moderns in many many different comparisons, until we get a result. I'll try to learn but I promise nothing (those specialist programs don't have easy to learn tutorials).
Updated again with consideration of the Dienekes comparisons. The European components are overly normal and emphasizes that both the "North European" (Balto-Russo-Finnic) and the "SW European" (Sardo-Basque) components are equally pre-Neolithic, but the 11% "East African" (and not "NW African") is not.
ReplyDeleteI mean it's not "normal", not "not pre-Neolithic", what it seems to be indeed.
Delete"But the big mystery is the East African component, which is distinct from the North West African (Moorish or Berber) one, which we might all have expected on first look. This East African component is highest among the Sandawe (68%), Somali (69%), Ethiopians (55%) and that's about it".
ReplyDeleteBut we do have East African Y-DNA through the Mediterranean. E1b1b1-M35 as a whole certainly looks East African in origin, and many of the E1b1b1a-V68 subgroups are also basically East African. In general E looks like an early traveller through the Mediterranean, presumably one of the first haplogroups to venture out into the wider Mediterranean islands.
Not likely to be the same thing, specially when that lineage has been found among Neolithic and not Epipaleolithic populations. But who knows?
DeleteThat would make better sense, IMO, IF the La Braña people would be a Neolithic people and not Epipaleolithic. They are borderline by chronology but we have no reason to think they were Neolithic, much less in such a peripheral mountain zone - so early.
You'd expect more Y-DNA A and B than you see, and also I agree with Maju that most of the E1b1b1 in Europe looks young and is in the upper branches of the phylogeny. Also, the African DNA in Europe has a SW v. SE split, with the SW having an affinity for Morocco and the SE having an affinity for NE Africa. So, a strong East African component in Iberia is very out of tune with modern populations.
DeleteIs it possible that this apparent East African admixture is a remnant of the OoA expansion from this region? I know it's odd, but supposedly these hunter-gatherers were closer in time and possible genetically too to the first HAM who left Africa.
ReplyDeleteThe components, zombies or otherwise, are affinity indicators. In this case, the sequence (or rather the composite of both sequences) is forced to align with a limited number of pre-selected components. The strange thing is that it chooses East Africa and not whatever else, and this is roughly consistent with its anomalous position in the global PCA plot.
DeleteBut what does it represent? Indeed it may be a residue of something that existed in the past and not anymore. But then East Africa is not more similar to the past than any other component (or should not be), so... perplex.
More analysis is needed.
The geographic spread in the H v. U5 ancient DNA map is suggestive of the notion of H as a possible Epipaleolithic expansion with a colony that is pushing U5 slightly inland.
ReplyDeleteAutosomal affinity of last to adopt farming and herding Northern Europeans with Mesolithic Iberians makes plenty of sense.
I'm not convinced that Dienkes k=7 or k=12 zombies have the resolution needed for the task, which may be more complicated especially in ancient DNA sets.
I also wouldn't be surprised if the "East African" affinity really reflects a much more wide ranging Mesolithic hunter-gather component that has been suppressed by post-LGM hunter-gatherer demographic expansion out of the Near East, as well as Neolithic and post-Neolithic population expansions, to which extant African hunter-gatherers are a best fit even if not a very good fit. Founder effects could also have greatly reduced post-LGM autosomal diversity among European hunter-gatherers who repopulated Europe over a fairly short time frame relative to the initial modern human population of Europe.
But what happens with all the likely (and in some cases as certain as it can be) H in Portugal and Morocco? What about nearly unmistakable H17'27 in Gravettian Sunghir? What about the possible other H (R-CRS) in UP Britain?
DeleteI think that there's too much speculation on this matter and that most likely H has been part of the UP genetic scene of Europe (and, maybe since Oranian, also North African and probably also of West Asia to some extent) since Gravettian or older times, more or less the same as U variants (mostly U5 and U4 but also some U*-CRS, now very rare, and U2 in at least one case).
However it is possible that some very common haplogroups like H1 and H3 expanded later, within intra-European Neolithic flows like Megalithism. For example the only pre-Neolithic H1 (H1b specifically) known is from Portugal, precisely where the Megalithic phenomenon was conceived and developed before its rather quick expansion though Atlantic Europe in the 4th and 3rd millennia BCE. H1 would fit very well with Megalithism, IMO.
"But what happens with all the likely (and in some cases as certain as it can be) H in Portugal and Morocco? What about nearly unmistakable H17'27 in Gravettian Sunghir? What about the possible other H (R-CRS) in UP Britain"
ReplyDeleteIt looks like there may have been population turnover in North Africa (from archaeology) ca. 16,000 BP. So, seeing H as an expansion out of the Middle East on both sides of the Mediterranean following the LGM but before the Neolithic, rather than out of the Iberian refugium, may be the best fit to the data. The dearth of H in Upper Paleolithic Northern Europe or at very old dates elsewhere, during which times and places U is predominant, suggests a less ancient H in Europe.
H1 and H3 expansion during megalithism has merit.
There's no such thing a an "Iberian refugium". Sometimes, not too knowledgeable authors use such name but they actually mean the Franco-Cantabrian region, which is only marginally "Iberian" and mostly Occitan/Aquitanian/South French.
DeleteAround the LGM however the true "Iberian province" of the Southern peninsula, became less marginal than usual and probably fed North Africa with fresh blood in the context of the (Gravettizing) Solutrean and its transmediterranean closely related counterpart: the Oranian (or Iberomaurusian) culture. This was c. 20 Ka ago (not c. 16 Ka. ago)
I have NO IDEA how can you fit, in archaeological terms your most strange idea of an "expansion out of the Middle East on both sides of the Mediterranean following the LGM but before the Neolithic". That has ZERO archaeological support.
Following the LGM, i.e. in the late UP, (South) Iberia lost again its relevance and Middle-West Europe (the Rhine) gained it again instead. But all the post-LGM process happen in or around France and have NO connections with West Asia (aka "the Middle East"). IF anything there could be flows TOWARDS West Asia, suggested by some scatter of Franco-Cantabrian-like rock art in Turkey and Egypt or some Eastern-European-like Gravettian influences in the Epipaleolithic of the Zagros. But all those are not well researched, so the connections are at best speculative (but would be Europe > West Asia and not the other way around in any case).
I strongly suggest that you re-read Bocquet-Appel's excellent paper on European archaeological demography (and in general improve your knowledge of European, West Asian and North African archaeology before you go around throwing weird ideas without any archaeological foundation).
I sincerely hope that your take on superstrings is more solid because what you're throwing here is indefensible from an archaeological viewpoint, what makes me doubt of your opinions on other matters where my knowledge is admittedly less solid.
hi Maju
ReplyDeletethere is another update from Dienekes..
"UPDATE III: In terms of the euro7 calculator, the results are: 89.6% Northwestern, 1.6% Southeastern, and 8.7% Far_Asian."
Closest to such apportions would be Irish or a range of modern Celtic/Germanic peoples - what is consistent with all we have seen so far. But the c. 10% East Asian (wasn't it African?) is still very perplexing.
DeleteI'd run a set of unsupervised Admixture comparisons if I could figure out how to insert a sequence and make up for the low coverage.
Updated again to include, among other stuff, the comment of Pedro Arias, a Cantabrian high-profile professor who argues that the funerary disposition of La Braña is anomalous for "Mesolithic" contexts, where true burial was the norm.
ReplyDeleteCorrection: Pablo Arias, not "Pedro".
DeleteBushmen!?!?!? That's crazy! Maybe there was a migration of Bushmen to the Iberia peninsula about 50 Ka? In any case, some other admixture-experiments have yielded about 2% of East african admixture (or whatever) in Spaniards.
ReplyDeleteI think it was you who suggested earlier it might mean some kind of generic Homo gibberish, maybe (who knows!) even Neanderthal remnants? The compass seems to be pointing to the root of either the Homo sapiens tree or at least the Eurasian root but it finds no correct reference so it wobbles.
DeleteSo maybe there is a 15-20% of really different ancestry from modern Europeans... and any other modern humans, at least their main components.
At this point, I'd compare by pairs or small groups with every other population and not anymore use zombies. I doubt these can say anything else.
It's very possible that the La Braña composite may produce its own distinct component and there the interesting thing may be to compare with others by means of Fst distances and such. The composites do not seem able to tell anything else.
"I think it was you who suggested earlier it might mean some kind of generic Homo gibberish, maybe (who knows!) even Neanderthal remnants?"
DeleteIn the Iberian peninsula? When did I say that? Maybe I was thinking in some odd types that live in Catalonia and look a bit like Bushmen or Asian? But definitely not neandertal, though I wonder why we aren't unable to see some neanderthal admixture in any Eurasian sample with these programs.
It seems these hunter-gatherers are mostly European and "something" else. It'd be very interesting to do more experiments to find it out, though I'm a bit suprised and disappointed, because it seems every experiment tells us a different history.
I did imagine Neanderthal admixture for a moment but then realized that such thing seems inconsistent with the info provided by the various Dienekes' comparisons, which wobble towards Asia AND AFRICA. Africans have no or much lower Neanderthal admixture than Europeans, so there's no reason for "the compass" to point in that direction.
DeleteBut only direct comparison can tell for sure. I'm guess that Lalueza has already done that or has it in mind...
Had a quick look at Dienekes' post on how the Braña sample "components" broke out with Dienekes calculator, and the HarappaWorld and MDLP5 calculators. Based on the differing results, I'd say that apart from a clearly consistent apportioning of the sample to Paleo Mediterranean and Western European components, the rest is noise.
ReplyDeleteHowever, it is notable that the MDLP5 calculator picks up a 16% East Eurasian component while HarappaWorld picks up a 6.7% Siberian component. In any case, below the 10% level, I'd say these calculators are just hunting through noise.
Not to say that there couldn't be something interesting going on, but you'd need more samples to even begin to figure it out.
But the noise is almost as consistent as the rest. If you want my bet an ancient OoA component now lost (seldom detected except maybe in the Slovenian HGDP sample).
DeleteI pondered Neanderthal admixture (now lost or heavily diluted) but then it should not point to Africa, even in the wobble.
We need a non-supervised Admixture series. But if you're waiting for me to do it first... it'll take a while because it implies reading and experimenting with technical gibberish... a major effort.
"We need a non-supervixed Admixture series"
DeleteHaven't looked closely at this, but perhaps more Braña samples or a more detailed SNP sequencing of these samples will be published.
"But the noise is almost as consistent as the rest"
What do you mean?
There are only two La Braña individuals: two funerary settings one beside each other. That's all. There's nothing else of archaeological relevance in that cave, which seems to have never been used for habitation (what is totally anomalous for Epipaleolithic contexts and would fit better with Neolithic ones in fact).
Delete"What do you mean?"
I mean that in all the various comparisons performed by Dienekes, some 9-17% has pointed to out of Europe ("the noise"). The exact direction and intensity of the "noise" compass varies but it's always a minor element in an otherwise fully European sequence (composite of two sequences in fact).
["European" in the sense of: modern European minus the West Asian (Neolithic) components].
The "noise" never points to an "obvious" origin like North Africa or West Asia or even India... but actually always outside of the West Eurasian genetic sphere: either East Asia or (East/South) Africa or both.
But East Asians and Africans are genetically quite distinct so the wobble only makes sense if it's saying something like "OoA lost component", like stuff I think can still be detected in Morocco and maybe even Slovenia. But while this interpretation is quite suggestive it would need to be fully demonstrated yet.
I did catch that there were only two La Braña individuals. So, if that is the case, perhaps new sequencing techniques or analysis can yield a less noisy result for these samples.
DeleteI've been wondering for a while as to the time depth capability of ADMIXTURE. It is able to pull out the Sandawe and the Hadza in Tishkoff's Africans runs. It is postulated that these two groups have been more or less isolated (until recently) from other Africans. The time depth of isolation is guessed at about 30,000 years.
I don't think Dienekes samples have enough definition to pick out archaic admixture at time depths equivalent to the Tishkoff paper. I'm also wondering whether his reference samples are set up to allow resolution of any African components.
To his credit, today, he did try alternative calculators.
In short, following the discussion here, I think people are putting too much in this result and not thinking about possible sources of error.
There are two individuals indeed. Dienekes made a composite out of them for technical reasons (very similar sequences - close relatives - and low SNP coverage).
DeleteAdmixture and similar algorithms have nothing to do with time: they detect genetic affinity/difference. Time is a factor in that divergence, of course, but so is isolation (endogamy), sometimes creating bizarre exceptions like we see among the Hadza or Tishkoff's Tunisian samples, which can be spotted and (optionally) discarded because of their strikingly low genetic diversity relative to neighbors.
"I'm also wondering whether his reference samples are set up to allow resolution of any African components".
Well, if African the "compass" should point to Africa all the time and not intermittently to Africa and Asia.
Without going into the specifics, Admixture, combined with Fst divergence time estimates, can be used to estimate divergence times and "time" the population splits you get when you increase the k value in an Admixture run.
DeleteI have used Fst distances to estimate time of divergence, yes. If you meant that, ok.
DeleteI do not think that the K values are of any use in that purpose, they depend not on time of divergence but on a combination of genetic distance and sample size ONLY (maybe you can add randomness in some borderline cases, but not time, as a factor).
The results keep changing because people use different datasets, the composition of a certain dataset, along with the specific SNPs used is what dictates results from model based algorithms like ADMIXTURE or STRUCTURE.
ReplyDeleteI still have not seen what the model based results of these specimen would be with a truly global dataset, one that includes all available public data points that were published minus any unverified private data points. In such a dataset and with the current SNPs that we have, Europeans and West Asians always come out to about 1/3 African and 2/3 East Asian when K=2, therefore these components become the building blocks from which further K or putative discrete ancestral populations are derived from.
From the original ancient DNA La Braña paper:
Delete"Furthermore, analyses of 1.34% and 0.53% of their nuclear genomes, containing about 50,000 and 20,000 ancestry informative SNPs ..."
They could only recover 50,000 and 20,000 SNPs, which is a low number of SNPs compared to the datasets that are now used for Admixture runs with modern populations. Also, we don't know the distribution of these SNPs across the genome. How random is its distribution?
Also, checking Dienekes' post, I notice he "pooled" the SNPs of the two individuals. Not sure that this matters, but it would be interesting to see how these individuals would appear if run individually.
50K and 20K SNPs maybe low for a spatially limited area, like say for instance Europe, but it is certainly enough resolution to tease out global substructure, quote from Sikora (2010):
Delete"We can conclude that the common set of 2841 SNPs genotyped is an appropriate tool to study population structure in African populations; in general, world-wide patterns are evident and robust when using a minimum of 1000 SNPs"
How much of these 50 and 20K SNPs would actually intersect with currently available datasets however I am not sure of, as I have not tried it my self.
With respect to the distribution, that is actually a good point, however once you intersect with a given dataset one can use PLINK to remove the intersected SNPs that are in high LD, which would even out the distribution of these SNPs across the genome, off-course this comes with a trade-off and that is a further loss in SNP density.
The ADMIXTURE software generally recommends ~10K markers for inter-continental studies (FST>.05), and for areas with FST <0.01, like Europe, ~100K markers.
The data is here: http://biologiaevolutiva.org/~clalueza/repository/
DeleteMy Admixture manipulation skills are too limited to do the comparisons however. So I would encourage others to do it or, alternatively, to tell me how-to (I'd put the computer time and learn something in exchange).
"The 2 specimens are more related to Northern Europeans, as they are closest to the British, who are Northern European??"
ReplyDeleteI don't think the British/Irish are "north european" in the context of Dionekes' components in this timeframe. In this timeframe i think they are atlantic-med with the north european component arriving later with the celt-germans.
I'd be most careful about equating Dienekes' zombies, with real populations. They are just (in origin) artifacts of an algorithm signifying genetic affinity. Dienekes favored those components but different sampling strategies and such may produce different results, for example a NW component, etc. He also decided to ignore some more specific components like the Basque one.
DeleteI strongly suggest caution on all that. I also doubt that Britain has ever incorporated so many immigrants in all post-Neolithic history (it'd be more than 50% if we accept what you say).
Incidentally this finding suggests that both the NE and the SW European components (affinities) were there (in León at least) before not just before the IE flows but even before Neolithic itself.
Agreed, i was just pointing out that in the context Dionekes uses them the "north european" component is more of a northeast european component that expanded west at some point.
Delete"I also doubt that Britain has ever incorporated so many immigrants in all post-Neolithic history (it'd be more than 50% if we accept what you say)."
Yes but i wonder if that was neccessary. There was a population explosion in northern europe in the middle ages with the invention of the heavy plow but only in certain areas i.e. heavily forested lowlands. In England that would have favoured the eastern half of the country.
Say for example if the west:east population ratio beforehand was 3:1 atlantic_med: northern european on a west to east cline then if the heavy plough created a three-fold population increase but only in the east then the ratio would become 1:1 afterwards.
I'm not saying that's what happened just that if one component of a population are disproportionately settled on land which *later on* with new technology becomes much more productive then their genetic signal could potentially be greatly multiplied.
These Leonese have both components in almost equal apportions, suggesting, if they are to be taken as reference (always due caution because they may be an exception rather than the rule), that both components were common at 50-50 apportions (roughly) in pre-Neolithic Western Europe.
DeleteYes, the context of this gets confusing. My view is Dionekes' northern component existed along the Atlantic in mixed form combined with the atlantic_med component but then also arrived at a later date from the northeast with the celts / saxons etc.
Delete"Bushmen!?!?!? That's crazy! Maybe there was a migration of Bushmen to the Iberia peninsula about 50 Ka? In any case, some other admixture-experiments have yielded about 2% of East african admixture (or whatever) in Spaniards."
ReplyDeleteit is most likely an "undiluted" endogenous North African component...similar to the one found in the Southern Moroccan sample at 15%...
the endogenous N.African has an African affinity,and that is why it was picked up by different African components rather the Northwest African component itself as one may expect...because the Northwest African is largely Eurasian....
so,the San did not migrate to the Iberian Peninsula...and neither did the Gedrosians ...after all an algorithm is just an algorithm .
http://blogs.discovermagazine.com/gnxp/2012/04/an-algorithm-is-just-an-algorithm/
Dalouh: we do not know. Only a direct comparison can demonstrate or deny what you say.
DeleteBesides the South Moroccan component has no more African than Eurasian affinity but it is equidistant at high Fst values 8very distant from all) - and that's why I think it's a remnant from OoA times.
There is another reason why I doubt it is the South Moroccan component: such thing could (should?) have pushed the compass towards the North African component but such thing never happened.
We'll see when the relevant unsupervised comparisons are made. My bet is a never before found, not any more existent, residual OoA component, different from that of Morocco.
"Bushmen!?!?!? That's crazy!"
ReplyDeleteWell, I don't know if there is a link but Y-DNA hg A1 is found in Great Britain:
http://news.bbc.co.uk/2/hi/uk_news/6293333.stm
Here is another independent analysis.
ReplyDeletehttp://fennoscandia.blogspot.no/2013/05/la-brana-and-modern-european-variation.html
Hi, Anders. I read it back in the day but I'm anything but satisfied by it. First of all, the method of PC/eigenvector analysis is extremely limited. I would much more prefer a sufficiently deep ADMIXTURE one because a bidimensional PCA is roughly equivalent to a K=3 ADMIXTURE depth (extremely shallow for most cases) and a 3D one (yours? - I do not understand well the graph) would only compare with K=4.
DeleteThen, I also find extremely hard to believe that there is any connection between these ancient early Neolithic or late Epipaleolithic (??? - the burial style looks Neolithic and the date is borderline) Iberians with the Finno-Ugrian populations of NE Europe, which have obviously undergone severe endogamy bottlenecks in "recent" times.
So I would say that your results are essentially a fluke or at the very least that they need of serious deeper analysis via ADMIXTURE and maybe IBD studies.
Still, I welcome all the possible info. But I'm very skeptic about your "findings", sincerely.
I also did a Finestructure-Chromopainter using the previous found substructures in Europe and the rest of the world from earlier 289k runs for the pruned 68K data. I have not published these but the La Brãna still shows a massive connection with the Saamis and Finns. I in that run I also checked the PCA. In Dimension 1 the La Braña clusted with the Saamis and the Finns. In Dimension 2 and 3 the same, but in Dimension 4 the La Braña move out towards some Basque together with two Saamis. So I have used different methods and got the same result. ADMIXTURE is not usefull in such small dataset without having large samples so I am currently working With STRUCTURE that is far more time consuming.
DeleteI have no clue about the archeological-cultural context of this burial, but isnt it possible in the borderline time that that cultural wave come before the genetic replacement at least some places?
DeleteIBD study of the La Braña is by the way difficult as we do not have diploid genomic data from them only "haploid" that is not possible to phase.
I have done a new analysis. This time using Chromopainter-Finestructure unlinked pipeline using "superindividuals" similar as "supervised" mode in ADMIXTURE.
Deletehttp://fennoscandia.blogspot.no/2013/06/la-brana-and-modern-world-variation.html
... "isnt it possible in the borderline time that that cultural wave come before the genetic replacement at least some places?"
DeleteSounds highly implausible to me. If the aboriginal population gets the cultural package in full (or almost) it's not anymore at disadvantage, so replacement should not happen at all after that.
But in any case, it's much more normal that the cultural transmission takes place first and foremost in practical elements of economic nature, like agriculture, cattle, pottery or innovative tools. Burial style is generally considered one of the most resistant cultural traditions because, while holding no practical value, it relates directly to the socio-cultural identity, to major traditions.
The issue is that La Braña 1 is set to rest in flexed position (typical of European Neolithic of Balcanic origins) in a cave not known to have been used before. European Paleolithic societies, in the West as in the East, tended to perform burials in extended position instead, often using ochre, and they are usually found near habitation sites, if not inside them. So I'm very perplex and would not discard these individuals belonging to an avant-guard Neolithic community, in a time that Mediterranean and SW Iberia at least was already Neolithic since some five centuries before, or arguably even earlier (the La Almagra pottery debate). Roughly at those dates of c. 7000 years ago is when Neolithic arrived to the Basque Country, according to the most recent studies, for example.
"I have done a new analysis".
It seems consistent with what you got earlier. The color scale is not very helpful anyhow, so I had the image imported to my PC and toyed with the brightness gradient to see better in that large dark-gray-to-black scale. It seems that the apparent Braña "influence" is greater into Finns2, No6, Saa1 and Lith (white when inverted) than among Basques and Croats (light blue when inverted).
I wish we had at least another data point, preferably a more reliably Paleolithic one, in order to be able to draw any realistic conclusions.
I have no knowledge about the prehistory in this area or the archeological cultural context, but I have worked much with genetics using different programs during the project and I know from experience that this analysis is not "fluked" as you call it. The software have been able to locate what populations in Europe these individuals genetically speaking are closest to of the modern populations today as with any other project participant I have.
DeleteRegarding the Finn "endagomy". I would believe that among any European hunter gatherer populations the individuals would be more similar to each other than any cosmopolitian population. This is because of the "natural" restrictions that apply to them as the populations can not grow bigger than the natural occuring food that is available. This means that they probably would need much bigger areas to feed the same population than a agricultural population, in effect it makes a roof of how big the population can be making genetic drift much stronger than in agricultural populations, and if food supplies varies because of natural cycles in the Food supply there would also be periodical bottlenecks.
DeleteThis phenomena of larger similarity among hunter gatherers you can also see among the Saamis, Siberians and Native American. If you look at the available ancient mitokondria data for European hunter gatherers you also there see dominance of certain haplogroups.
Of course: hunter-gatherers usually need bigger areas than agriculturalists, although it may not be too different than free range herders, such as were Finno-Ugric peoples until relatively recent times (or are still the Saami and others).
DeleteAnyhow even foragers there seem to have diverse levels of effectiveness in land area exploitation. For example we see in Europe a clear signal of demographic explosion in Magdalenian times, surely caused by the improvement and intensification of the exploitation of aquatic resources (fish, seals, whales, other seafood) after the LGM. This kind of intensive aquatic-based foraging is known to have produced other very dense and affluent forager societies for example in NW North America. So it's not a simple matter.
Also, even in the LGM, SW Europe (for example the Franco-Cantabrian region) was warmer than today's subarctic Europe. This is evidenced for example in the limited presence of reindeer and abundance of other animals of warmer latitudes like deer, bison, auroch, horse, etc. Everything else equal, the Franco-Cantabrian Region could feed many more people in the Ice Age than a similarly sized (or even larger) area of the subarctic Finno-Ugrian belt in post-glacial times before the spread of agriculture.
Therefore it is not such a simple comparison as you seem to think.
Also the likely larger similarity (endogamy) within hunter-gatherers must be considered always only within populations. It's probable that the late UP hunter-gatherers of Aquitaine and those of Ukraine were not related except for remote ancestry, even geographically closer regions like Italy were in the same situation (no Solutrean nor Magdalenian means no or very limited genetic flow with SW and Central Europe).
Less clear but also possible is that, for example, the various Magdalenian populations of most of Iberia (Iberian region), the Franco-Cantabrian region and the Central European region were only weakly related among them. After all the time-line of Magdalenian is of some 7000 years (even more if we consider Magdalenian-derived Epiplaeolithic), about the same time as from the beginnings of European Neolithic to present day, allowing for many possible divergence and processes (never mind the pre-Magdalenian substrate).
So I would not assume even for a second that all European hunter-gatherers were closely related upon arrival of Neolithic: diversity must have been very important. Within regions (and more so within subregions) instead is very likely that they did experience some intense endogamy, depending also on the actual population densities, which were always quite greater around Dordogne (~Perigord).
If you look to Siberian and Native American populations and run f.ex in PLINK a PI_HAT analysis it will basically tell you they are all cousins that is sharing up to 25% even when their real relatedness is far beyond counsins level by many thousands years. However if you do a haplotype based analysis for Siberians and Native Americans that is able to count all the unique mutations each individual have (like in BEAGLE) they will show an unique mutation rate occurance close to Fennoscandians. Any closely related individuals that is actual cousins will have a strongly reduced occurence of these individual unique mutations. What I hear mean by unique mutations are mutations not seen in any other individual in a population panel in any given haplotype. In Europe Southern populations typically have far higher unique mutation for each individual than a Northern populations.
DeleteSo in this respect I have no problem to see that the ancient European populations especially soon after the ice reatreat covered large distances was more similar to each other.
I do not understand what you mean. I was like nodding all the time when reading your first paragraph and then, when I read the second, like: "how did he link both things?!"
DeleteI mean: you acknowledge that Far North Eurasian populations show characteristics typical of extreme population history, while Southern Europeans do not. However Southern Europeans are equally descending from Paleolithic hunter-gatherers as the Northern European endogamous peoples (I don't care here if it is European or West Asian Paleolithic hunter-gatherers or a mixture of both: West Asians were also hunter-gatherers until a few millennia ago, you know). So it is evident that people who descend from Paleolithic hunter-gatherers (everybody) can have and usually have a richer genetic diversity than Far-Northern Eurasians (logically, as their regions' warmer climates, excepted arid ones, allowed for much greater population densities in the past, as in the present).
So ancient West Eurasian populations must have been somewhat diverse, at least with differences from region to region, from population to population, and in some cases at least (those areas with greater densities such as that Franco-Cantabrian region), also inside the populations.
Whatever the case it's irrelevant how you want to imagine it because what you imagine is not based on material evidence: Dniepr-Don, Italian, and Doggerland Paleolithic peoples had very limited contact, if any at all, with the SW peoples of Solutrean and later Magdalenian and Epi-Magdalenian culture (i.e. since the LGM). Otherwise we'd see the result of those contacts in the material archaeological record, and we do not.
In fact it's much more likely that some of these peoples had meaningful contacts with North Africa and West Asia than among them, as suggested by some material archaeological elements like rock art and also some techno-cultural flows. Sadly the archaeology of North Africa and West Asia is often even less known to most people than European one (which is surprisingly misunderstood by way too many people attempting to reconstruct the past from the perspective of population genetics, particularly), this creates a "here there be dragons" effect that is interpreted by some as a blank check to fill in the "blanks" (no blanks in fact in most cases, just ignorance) in whatever way they want. This has given rise to myths like the "Iberian refuge", the "Italian refuge", the "Balcan refuge", etc. which hold no ground whatsoever. Another such myth is imagining Mesolithic/Early Neolithic West Asians as some sort of Mary Sue that would have somehow all the imaginary attributes that are denied to European ones: quasi-infinite diversity, quasi-infintite capacity for boundless expansion, total disconnection with Late Paleolithic Europe, etc.
...
...
DeleteBut, hey, look: a few millennia before Europe was Neolithized, West Asians were also hunter-gatherers with very similar same kinds of constraints or attributes as those found in Europe. Look: West Asia has some 6 million km² (more than half of them very arid, also in the Ice Age), while Europe has some 10 million km² (of which only a third or so was covered by ice at the LGM). So being generous they are very much comparable regions, and being strict, Europe has some advantages. The area of Magdalenian culture alone has some 2 million km², something no other West Eurasian or Mediterranean Late UP culture had other than maybe North African Oranian (Iberomaurusian). I mean, Turkey, for example, which covers much more area than any West Asian Late UP culture, has 800,000 km², just slightly more than France.
So we can't argue, as you seem to imply, that, somehow, miraculously, West Asians had more greater genetic diversity than Europeans at the onset of Neolithic, except maybe because of the fact that the original colonization of WEA arrived via West Asia and therefore left some greater basal diversity over there.
Nor we can argue, unless material archaeology supports it somehow, that people were migrating thousands of kilometers across Europe (why Europe only and not all West Eurasia: there was no Bosporus back then?) making miraculously again and against the material evidence, Ukranians and Iberians the same Paleolithic super-population even if their cultures diverged in the Gravettian period, c. 30,000 years ago.
New updated analysis. http://fennoscandia.blogspot.no/2013/06/la-brana-and-saamis-ii.html
DeleteI have by the way today seen Vadim Verench data on the La Braña and the ancient Gotlanders. They where quite similar autosomal speaking.
Deletehttp://verenich.wordpress.com/2013/06/13/практические-рекомендации-по-работе-3/
Correct me if I read wrong but the heatmaps show that the "hottest" samples are Norwegians, not Saami. This would make better sense because the bulk of Norwegian and other Scandinavian ancestry should be from Western Europe, while the bulk of Finnic (incl. Saami) ancestry should be from Eastern Europe.
DeleteI'm reading from left to right, i.e. in the rows. If I read the columns, then La Braña is closest to Norwegians and Danes in one graph and to Norwegians again in the other. I can't understand why you insist on Saami affinity with that data.
It is very hard to fathom how the La Braña people, be them Epipaleolithic or Neolithic, could have its closest connections with the Eastern European ancient substrate, probably manifest in Finnic peoples today to large extent. A Western European affinity makes at least some better sense, even if the possible Paleolithic ancestry of that area (Scandinavia and other North Sea peoples) should be more from Doggerland's Ahrensburgian-Hamburgian-Maglemosean than from SW and Central European (Epi-)Magdalenian.
If you mean NO6 and NO7 these are Norwegian-Saami heavy mixed individuals. You can see it very clearly in the normal 289k Europe analysis and in the normal 289k Fennoscandia only analysis. SWE7 is a South-Saami individual. However it doesnt matter as if you remove these mixed individuals the clustering still hold vs the Saami.
DeleteYou cant by the way only read single blocks like that. If you compare the whole painting of the DK1 (The Danish) with the whole painting of the La Braña you see very clearly the difference between them even they do appear to share something in common. The location on the tree of the Danish is not weird at all and make perfect sense so does it for the La Braña individual given their paintings in the heatmap.
DeleteYes, I meant NO7, NO6 and DK1. I also notice now the SA1-2 and SA3 in one of the graphs (missed earlier). So I see now better your viewpoint, based on your results.
DeleteHowever I am still in the disbelief zone because, in principle, there is absolutely no reason to imagine any particular relationship between this composite "individual" (BRA is actually made up of limited sequences from the two La Braña burials) and Saami, who are not only of an almost totally different background but also have experiences major bottlenecks in their population history, what should drift them very far apart from all other populations (as we can usually see in other studies and analysis).
So let me tentatively criticize the methodology: AFAIK, heatmaps are related to TreeMix software/algorithm which is known to produce very strange results (for example, but I've seen many others). Heatmaps themselves also do often enough (for example).
PCA is also of limited interest (equivalent to ADMIXTURE's K=3 or so) and highly susceptible to sample size.
So I have to remain skeptic because neither the methodology is reliable, IMO, nor the apparent results seem to make any sense.
PS- The Verench analysis on the other side is interesting and coherent to what we have seen in other ADMIXTURE-like analysis. It is interesting that even Ötzi clusters very closely to Swede TBK.
DeleteCan you explain me what do exactly TBK and PWC mean? I can't read Russian. I'm assuming that TBK is the same as the German acronym TRBK, i.e. Funnelbeaker culture (macro-culture), probably with Megalithism, but I do not know enough and may well be wrong.
As I understand the Chromopainter-Finestructure science papers so far they are some of the best clustering software available. I cant rememmber read anything about TreeMix in their papers. As I understand it they are based on coancestry.
DeleteHere is a link to their Chromopainter-Finestructure site:
http://www.maths.bris.ac.uk/~madjl/finestructure/chromopainter_info.html
Here is also a link to a presentation:
http://www.maths.bris.ac.uk/~madjl/Lawson%20finestructure%20Popgroup%202012.pdf
"You cant by the way only read single blocks like that".
DeleteCan't I? Isn't the heatmap comparing sample with sample on one-to-one basis?
"If you compare the whole painting of the DK1 (The Danish) with the whole painting of the La Braña you see very clearly the difference between them even they do appear to share something in common".
Beats me, honestly. How does it matter how they relate to third samples to their one-to-one comparison? This would only mean to make any sense if you are looking at the heatmap with "PCA eyes", so to say. But PCA is a very weak tool, so I won't do that.
"The location on the tree of the Danish is not weird at all and make perfect sense so does it for the La Braña individual given their paintings in the heatmap".
Autosomal DNA and phylogenetic trees are almost antonyms. By definition autosomal DNA mixes and remixes every generation, hence trees can only be at best a very rough approximation and, at worst, a meaningless source of confusion.
So I do look at the heatmaps as one-on-one comparisons and I do not think this is wrong at all.
TBK is Funnelbeaker culture
DeletePWC is Pitted ware culture
http://en.wikipedia.org/wiki/Funnelbeaker_culture
https://en.wikipedia.org/wiki/Pitted_Ware_culture
I take notice of what you say about Chromopainter. I guess I was wrong in associating it with TreeMix. Thanks for the clarification.
DeleteI'm reading the PDF as I write this and I notice something quite interesting in the European pops. example: Basques cluster best in both reading directions with the French, however the tree places them along with Sardinians, a clearly "colder" relative. Exactly the same happens to the French (and Orcadians): they are located in the tree along Tuscans and other Italians, when it is clear that they share a quite weaker "heat" (orange) than with Basques (red).
Looking at the fine detail graph a few pages later, we see part of why this happens: French at fine detail, tend to cluster overall (purple shades) similarly with Russians, Italians, Tuscans and Basques. Orcadians (a tiny population of complex ancestry), who are the best match with French, cluster worse with Basques and better instead with Italian/Tuscans and Russians. So the tree we see in the first graph is an artifact of the excessive weight of Orcadians. The Basque-Sardinian branch is not anything real but a catch-all "others" grouping (it has no other explanation: Basques still cluster much better with the French, and even the Orcadians, than with Sardinians).
So it would seem that the tree produced by Chromopainter is quite irrelevant because of its high susceptibility to sampling bias. The one-on-one comparison seems to be the most informative, assuming that the algorithm works well.
Still I am puzzled by what seems to be some Chromopainter results, as mentioned before. In Pugach 2013, fig. 4 B-D, they get some very shocking "heat" results, with Dravidians very strongly clustering (blue) with North Han, etc. I cannot imagine why these results.
Your interpretation is too simple. You need here also to take into account the assymetry between chunkcounts received and chunkcount donated.
DeleteAs you can see here from the example you found the French "like" to paint their chromosomes with Basque chromosomes, but the Basque do not like as much to paint they chromosomes with French chromosomes. This means there is a considerable asymmetry meaning larger genetic distance between French and Basque haplotypes. You dont see the same asymmetry between Basque and Sardinians even the number of chunkcounts are smaller.
You also see small asymmetry between the French and Orcadians, this means their haplotypes are very similar but the French haplotypes are not so similar to the Basque haplotypes.
I'm learning some stuff, so I do appreciate this discussion, Anders. Thanks for your insistence.
DeleteSo, according to what you say:
"the French "like" to paint their chromosomes with Basque chromosomes, but the Basque do not like as much to paint they chromosomes with French chromosomes".
... then Basques are (somehow, loosely speaking - I know that "we are not our ancestors") important "ancestors" (donors) to the French (sure, you only have to look at their faces in most cases) and, to lesser extent, to other populations like Orcadians, Sardinians, Italians, etc. Then French/Orcadians are the second most important "ancestors" (donors) in general, Italians and Russians. Etc.
However this still not reflected in the tree. For example, Basques are always more donor than receptor in comparison to all other populations, what should make them hang directly from the root. So I really do not understand how the tree is made at all, nor I appreciate that it would have any correlation with the ancestral reality depicted by the color coding.
I understand that top populations are donors and left side ones receptors, right? Else it'd be all the way around but that's what I deduce from your comments.
"This means there is a considerable asymmetry meaning larger genetic distance between French and Basque haplotypes".
It should not be that way because even if the French are less important donors to Basques, they are still more important than Sardinians. I.e. Basques are still more French than Sardinian, even if French are more Basque than Basques are French. In fact Sardinians are more donor ("ancestral") to the French than to the Basques, so, really, the groupings made in the tree make no sense whatsoever.
At least that's how I see it, honestly.
PS- It is quite intriguing to me that, according to that Chromopainter graph, Basques are more donor (i.e. ancestral) to all other populations in Europe, including Adygei, than receptor (i.e. descendant). The same seems true for the French, with the only exceptions of Russians and Orcadians (which are more balanced). Doesn't this seems to strongly support the Paleolithic expansion model from, primarily, the Franco-Cantabrian region?
DeleteActually this is the same trait as Finns and Saami have versus many other populations in Chromopainter asymmetry. They prefer painting their own chromosomes less with the chromosomes of other populations than they donate to other populations.
DeleteThe Africans vs non-Africans however is always painting more with non-African populations than donating to non-African populations.
Majo wrote: "However this still not reflected in the tree. For example, Basques are always more donor than receptor in comparison to all other populations, what should make them hang directly from the root. So I really do not understand how the tree is made at all, nor I appreciate that it would have any correlation with the ancestral reality depicted by the color coding"
So basically if you have large asymmetry whatever direction - less painting and more donating, or more painting and less donating the most extreme population of these will branch out earlier from the main larger tree they belong to.
DeleteI do not understand. Basque always branch out early from the larger group they belong to unless they are in panels that include Spanish or/and Iberian samples.
DeleteMaju wrote: "However this still not reflected in the tree. For example, Basques are always more donor than receptor in comparison to all other populations, what should make them hang directly from the root. So I really do not understand how the tree is made at all, nor I appreciate that it would have any correlation with the ancestral reality depicted by the color coding"
So do I. I appreciate your critical approach.
DeleteMaju wrote: "I'm learning some stuff, so I do appreciate this discussion, Anders. Thanks for your insistence."
"I do not understand. Basque always branch out early from the larger group they belong to unless they are in panels that include Spanish or/and Iberian samples".
DeleteFirst of all, they should not be grouped with Sardinians (why?).
Second, their affinity to other Western Europeans is clearly larger than with any other peoples, and even the other West Europeans (French, Orcadians) are closer to Basques in either sense of the comparison than to MOST other peoples (at least). So, as I understand it, this means that the cluster A includes Basques, French and Orcadians, something that is obvious in the one-on-one comparisons (and also in haploid DNA).
This does not happen (so strongly) in Admixture however, so there seems to be something strange in the way the algorithms deal with actual dominant affinity.
I really do not understand why would actual affinity values (each 1:1 comparison) be subtracted, if anything they should be averaged, right?
You should view my my latest 289k SNP Euro analyss both for number of chunkcounts CC and for total length shard in cM.
DeleteIn the CC the Sardinians cluster with Italians but still share high number of CC, and both Sardinians and Basque is in the same tree neighbourhood.
https://sites.google.com/site/fennobga/CCEuroAggr24042013.png
In the CL the Sardinians cluster with Italians but still share high number of CL wit Basque, and both Sardinians and Basque is in the same tree neighbourhood.
https://sites.google.com/site/fennobga/CLEuroAggr24042013.png
In the World CC analysis the Sardinians and Basque is sharing close branches together with the Spanish, Italians and some French. This is due to the fact that local substructure is less seen in World analysis compared to the Europe analysis. Still Basque and Sardinians in the same neightbourhood.
https://sites.google.com/site/fennobga/CCAggrWorld240413.png
I have some more populations than the 650k HGDP study in that Chromopainter presentation. However they have much more markers and much higher resolution so its fully possible they are closer to the truth than I am in my analysis.
So yes it make sense to me that Basque and Sardinians is in the same part of the tree as that presentation shows.
Also I did soem time ago do a ADMIXTURE analysis using only 22k but a huge number of samples. Basque was 31.1% Med and 64.6 % North Euro, and Sardinians 50.2 % and 46.5 %. PCA plots have also placed Basque and Sardinians in the same neighbourhood.
Maju wrote "First of all, they should not be grouped with Sardinians (why?)"
I do not understand the difference between CC and CL but, in any case, while some of the results of your trees are somewhat more reasonable (Western Europeans tend to cluster together, at least much more than in the other example), there are still many apparent inconsistencies.
DeleteA very clear one is Europeans clustering primarily with Africans, while Native Americans and East Asians are the first to diverge. So this shows that the tree algorithm, at the very least, is highly susceptible to low relevance elements, as may be recent minor admixture or other artifacts.
Even the heatmap produces very strange results, like Europeans being important donors to Africans, including some very isolate and distinctive peoples like the Mbuti or the San. I really think that inter-continental application evidences the limitations and output errors of this algorithm.
"Also I did soem time ago do a ADMIXTURE analysis using only 22k but a huge number of samples. Basque was 31.1% Med and 64.6 % North Euro, and Sardinians 50.2 % and 46.5 %. PCA plots have also placed Basque and Sardinians in the same neighbourhood".
I'm more and more feeling that using Sardinians as a reference population (especially in large numbers) in Admixture/Structure really distorts the results, as does using Finnic peoples, because both groups are heavily "drifted" or "bottlenecked". Basques instead do not show such intense endogamy, although some less important one yes.
I do not know what exact what map you refer to, but anyway my interest is in European and especially Fennoscandian ancestries so I usually use few samples from other places. It probably affect somewhat the structuring of these populations, f.ex the Africans have tendencies to branch with South-Asia, Oceania, East-Asia populations.
DeleteYou should instead read the 650k HGDP based Chromopainter-Finestructure run done by the Authors. It is far more reliable when it comes to populations outside Europe.
http://www.maths.bris.ac.uk/~madjl/finestructure/Clear_WorldCopymap1a.png
Maju wrote "A very clear one is Europeans clustering primarily with Africans, while Native Americans and East Asians are the first to diverge. So this shows that the tree algorithm, at the very least, is highly susceptible to low relevance elements, as may be recent minor admixture or other artifacts."
CC - Number of chunkcounts - that is number of identical and related segments.
DeleteCL - ChunkLength - the total amount shared of CC measured in cM (centimorgans).
Maju wrote: "I do not understand the difference between CC and CL"
Again you need to see the large asymmetry between painted and donated and tree clustering. The Africans have far more diverse haplotypes than non-Africans. Therefore they find distant but related haplotypes outside Africa. However non-Africans is far less diverse so they dont paint with African haplotypes but instead paint with haplotypes from populations outside Africa. This indicate large divergence time. In addition I use very few African samples so this scale escalate the copying. Its far more balanced in the authors study. I also scale Down the colors to make it easier to see smaller differences within Europe.
DeleteMaju wrote: "Even the heatmap produces very strange results, like Europeans being important donors to Africans, including some very isolate and distinctive peoples like the Mbuti or the San. I really think that inter-continental application evidences the limitations and output errors of this algorithm."
"I do not know what exact what map you refer to".
DeleteThe last one you linked to. AFAIK the only one to include non-Europeans.
"... but anyway my interest is in European and especially Fennoscandian ancestries so I usually use few samples from other places".
I can understand that but for the purpose of testing the method, intercontinental comparison seems very useful, because we know quite well how these various macro-populations relate to each other.
"Again you need to see the large asymmetry between painted and donated and tree clustering. The Africans have far more diverse haplotypes than non-Africans. Therefore they find distant but related haplotypes outside Africa".
But should not they be for that very reason massive donors (and almost never receptors)?
"However non-Africans is far less diverse so they dont paint with African haplotypes but instead paint with haplotypes from populations outside Africa".
This is totally counterintuitive. Non-Africans are "descendants" from Africans in some 97%, so they should "logically" (my logic anyhow) be massively "painted" by African sequences.
And in any case Africans should NOT be "painted" by European sequences. This last especially makes absolutely no sense because (excepted some admixed populations like Ethiopians) West Eurasians are not ancestral to them at all.
Yes, buu if you follow down the trail you see its because of the Middle Easters and South Asians. Many of them appear to have African like influence. That trail goes all the way to the Papuans and Melanasians. Opposite to most Europeans they actually do paint their chromosomes with the African chromosomes. Because the African sample size is small they go down that trail as there is clsoer relationship there.
DeleteMaju wrote: "The last one you linked to. AFAIK the only one to include non-Europeans."
Actually using just a few individuals from non-European populations capture quite well for my purposes even the African tree structure is not Perfect because of small sample size. However I know that and it has been shown in the authors study. I have nothing more to prove there.
DeleteMaju wrote: "I can understand that but for the purpose of testing the method, intercontinental comparison seems very useful, because we know quite well how these various macro-populations relate to each other."
If you investigate the output files from a Chromopainter run you can see many things. One of the output files it makes is a file refer to the number of mutations observed in the painting and donoring between individuals.
DeleteExample SWE2 vs San.
The San paint 16.5 cM of his chromosome with SWE2. SWE2 only donate 4.2 cM to the San. The counted mutations during paint is 25.5 and donate is 1.4. If we calculate the number of mutations per cM: paint 1.6 and donate 0.28. The calculation for San vs a Pygmy: paint 4.14 and donate 3.96.
As we can see the San easily find segments to paint in the Swede but the Swede have a hard time finding segments to paint in the San because the Swede simply lack the genetic diversity in their own haplotypes compared to the San.
As I understand it non-Africans came out of Africa as a narrow subset of the existing African population and the following New founding to the rest of the world narrowed it even further. They have of course also developed further their haplotypes making it even harder to find copying.
Maju wrote: "This is totally counterintuitive. Non-Africans are "descendants" from Africans in some 97%, so they should "logically" (my logic anyhow) be massively "painted" by African sequences."
"However I know that and it has been shown in the authors study."
DeleteIs that study available? If so, please, link.
"The San paint 16.5 cM of his chromosome with SWE2. SWE2 only donate 4.2 cM to the San. The counted mutations during paint is 25.5 and donate is 1.4. If we calculate the number of mutations per cM: paint 1.6 and donate 0.28. The calculation for San vs a Pygmy: paint 4.14 and donate 3.96.".
This is overly complicated. I really have more than a hard time understanding what you say.
http://www.maths.bris.ac.uk/~madjl/finestructure/Clear_WorldCopymap1a.png
Deletehttp://www.plosgenetics.org/article/info%3Adoi%2F10.1371%2Fjournal.pgen.1002453
http://www.maths.bris.ac.uk/~madjl/finestructure/LawsonEtAl2012-Finestructure-SupportingInformation.pdf
Maju wrote "Is that study available? If so, please, link."
First it shows that the number of mutations between any given segment between a San and a Pygmy is far higher than vs the Swede.
DeleteSecond it shows that due to the higher diversity of the Sam he can paint his chromomosomes with the Swede haplotypes, however the Swede with a very narrow haplotype only find few matches in the far more diverse San. Instead he Paint his haplotypes found in the non-African populations he have more in common with.
Maju wrote: "This is overly complicated. I really have more than a hard time understanding what you say. "
I've been reading and re-reading the Lawson 2012 paper and I am pretty sure now that I as doing it correctly all the time: what matters is how any individual or population (receptor) is decomposed in terms of the other indiv./populations studied (donors). There's no or very little more mystery to Chromopainter: it's rather straightforward:
Delete"... every individual is ultimately reconstructed in terms of all the other individuals".
I.e.: the receptor is made to be a sum of fragments (chunks) of the other individuals (donors). Same when dealing with populations.
So I have been reading CP well all along.
Another thing is tree construction, for which they are inspired by an actually unworkable misuse of ADMIXTURE-like algorithms, by which the set is divided by force assignation at K-levels in sequential order. Theirs is similar (less important changes) and therefore equally unreliable.
So I was doing the correct thing when mostly ignoring the tree as unreliable or merely "decorative".
Less important issues:
Most graphs do not explicitly indicate which list of populations represent donors and which are the receptors. I understand from context that, unless otherwise indicated columns (top or bottom list) indicate donors and rows (side list) are receptors.
In some graphs the population influences are renormalized so each continental region has its own independent scale. Again I miss that this is not always explicitly indicated because it can induce confusion.
Thanks for the links and your patience, Anders.
I have been working with this software for quite some time now and know it therefore from the practical point of view.
DeleteIts quite sensitive software and can even detect very subtle ancestry differences in f.ex my Fennoscandian local analysis when doing the linked 289k analysis. Especially challenging has it been to find the South-Saami minority ancestry in Scandinavians and even differantiate it from Finn minority ancestry in Scandinavians. No software to date have been able to do so.
In 68k unlinked mode the resolution will of course becomme lower but it has been able to infer the population structures I have seen in the 289k linked analysis.
So to try bury the results in theorical issues about supposedly faulty methodologies in the software doesnt work, as long as the practical applications have given a result that fully match the actual ancestries of the individuals in the project.
The La Braña's closest cousins of today genetically speaking of the available population panel is without much doubt first the Saamis and then Finns, and that have also been confirmed by another quite skilled genome blogger Vadim Verenich.
Maju wrote: "So I was doing the correct thing when mostly ignoring the tree as unreliable or merely "decorative"
Actually if anyone want to investigate the driving SNP's in PLINK's MDS plots it fully possible to find them using the -assoc command.
DeleteIn the dimension 1 or X-asis of this plot you can find some of the most significant SNPs and find how the frequencies distribute among the populations.
http://fennoscandia.blogspot.no/2013/05/la-brana-and-modern-european-variation.html
Like rs12219610 on chromosome 10. It has a C and T allele. The C allele is in my samples used mostly found in Fennoscandia and North-Eastern Europe but is almost absent in Western Europe.
The C allele is absent among the Basque sample I have used, its absent among the Spanish sample I have used but it is found once in 1000G iberian sample I have used.
The La Braña have the C allele. So this way one can continue with several thousand more significant SNP's using this method and you will end up with the conclusion that the positioning of the La Braña in sum actually should be placed where it is.
This comment has been removed by the author.
ReplyDeleteThis comment has been removed by the author.
ReplyDeleteAnders, you are probably aware of this data already, but here you have the Norwegian and Finnish Saami mtDNA haplogroup V and U5b frequencies and Swedish and Finnish Saami highest Y dna frequencies:
ReplyDeleteNorwegian Saami: U5b 57%, V 33%
Finnish Saami: U5b 41%, V 38%
Swedish Saami: N1c 37%, I1 31%, R1a 20%
Finnish Saami: N1c 55%, I1 41%, R1a 3%
About half of Saami R1a should be R1a (xR1a1), and also rare I* has been detected in Saami according to Wikipedia.
My hunch is that both U5b and V are very old in Europe, and probably also R1a (xR1a1) (found in Western Europe) and I*.
Kristiina. Yes as I understand Tambets and other studies the variation of those haplogroups points origin towards Southern and Southwestern Europe. It that regard the result my autosomal analysis seem to make sense.
DeleteKriistina, I read in Tambets et al 2004:
Delete"Soon after the beginning of the retreat of the ice sheets covering the area, in the 8th–10th millennia before present (BP), populations of hunters and fishermen, the producers of the Mesolithic Komsa and Fosna-Hensbacka cultures, inhabited the coastal region of Scandinavia, extending well into Finland and to the Kola Peninsula (Kozlowski and Bandi, Nygaard).
The linguistic affiliation of these pioneer settlers of the north is largely unknown, but it has been suggested that they are the descendants of the Ahrensburgian population, which migrated toward the north from western Europe, along the Atlantic coast of Norway. It has been proposed that they might have been the ancestors of the present-day Saami (e.g., Sumkin).
Another presumably important component in the postglacial recolonization of northern Fennoscandia came from the east, via Karelia and Finland. It has been associated with the movement of Mesolithic populations, carriers of post-Swiderian cultures, to the north. Starting in the Neolithic period, the northern population came into contact with tribes of territories lying to the south (e.g., Sumkin). Thus, according to archeological data, the present-day Saami population might have been shaped in different times both by the eastern and western influences."
"This genetics-based reconstruction (fig. 4) is in agreement with the reconstruction of the spread of Ahrensburgian and Swiderian Mesolithic technologies in northern Europe, linking it with population expansion that can be likely traced back to the post–Last Glacial Maximum recolonization of the European north (Torroni et al.
2001; Tambets et al. 2003). The results also stress that the grouping of populations according to language families should be used exclusively only in a linguistic context."
So you think that Bra is Eastern European of Western European origins (Solutreo-Swiderian)? Fine. I have the right and duty to remain skeptic however, notably because it's all so complicated and thinly painted that everything remains possible, even that.
DeleteIncidentally Swiderian has been argued to be derived from Solutrean (of which there was a branch in Hungary). Similarly I have argued that the (NW) African admixture in Iberia may have arrived to this region within the Solutrean period, the only reasonable explanation to explain its extension through Asturias and Cantabria but not the Basque Country (yeah, it does fit with the UP archaeological record, regardless of whether it is correct or not).
Can this Bra-Saa link you seem to perceive so clearly be of Solutrean (or Gravetto-Solutrean) origin? That would still allow for further changes to happen with Magdalenian and Epipaleolithic in most areas.
Archeology, linguistic and genetic are all different disciplines, therefore "inconsistencies" between them doesn't invalidate any of their results as they describe very different things.
DeleteImagine future archaeologists find what they call the "T-shirt" culture over large distances of Europe. They find T-shirts in Basque, Scandinavians, Finns and Saami burials but when they analyse their physical remains they get a totally inconsistent results with the archeological T-shirts. As they all weared T-shirts they also should have similar genomes, but they didn't.
What of these two seemingly inconsistent results are "fluked"?
You can't put in the same plane a hard science like Archaeology with soft ones like linguistics or population genetics (yes, genetics is a soft science, scoring similar as economics and well under ecology or agricultural science, not to mention space science). It will depend on the particular facts or claims being compared but in general, what Archaeology produces are facts with high value as evidence while statistical results in population genetics or linguistics are way too often just more or less educated hunches.
Delete"What of these two seemingly inconsistent results are "fluked"?"
A fluke (noun) is "A lucky or improbable occurrence, with the implication that the occurrence could not be repeated". As verb: "To obtain a successful outcome by pure chance". (Wikitionary)
I suspect that there's such element in your results because it's very difficult to explain them in any other way, but it's maybe better described as "artifact" than "fluke". Very possibly, the same that the Finnic ultra-homogeneous gene pool causes artifacts in ADMIXTURE and other statistical comparison tools (same for other populations with low heterogozygosity like Ashkenazim, Kalash, etc.), it also happens in ChromoPainter. The fact that you heavily load your samples with rare populations from the Far North, often very homozygous, doesn't help but to aggravate the problem.
"As they all weared T-shirts they also should have similar genomes..."
Nobody ever claimed that. Serious archaeological reconstruction of cultural realities is not based on single elements but on all kind of evidence. For example most prehistorians agree that Megalithism or Bell Beaker do not represent any single population but wider phenomenons across otherwise locally rooted cultures, similarly the Cardium-Impressed Pottery culture is documented to have some true colonies and also, very often, locally rooted communities which only adopt from CIP the new elements such as pottery and agriculture but not for example the toolkit, typically of local Epipaleolithic roots.
But that does not mean that well understood cultural units are meaningless either and in the event of Epipaleolithic colonization of (previously frozen) Northern Europe, which was the point I was arguing, cultural affinity offers us important clues on possible origins of the various populations. Just denying these evidences is not helpful at all, especially if you don't offer alternative clues of the same kind, such as evidence of locally rooted gradual or partial cultural change.
Thats not a problem in Chromopainter-Finestructure pipeline when using superindividuals for all populations except the La Braña. As you can see here the structure make very much sense even the sample sizes are very uneven. The heatmap structure also form nice boxes suggesting the structure is quite right.
Deletehttp://fennoscandia.blogspot.no/2013/06/la-brana-and-saamis-ii.html
Maju wrote:
"The fact that you heavily load your samples with rare populations from the Far North, often very homozygous, doesn't help but to aggravate the problem."
I am not sure what this conclusion is based on but they are welcome to check out the subtle ancestry inference I have infered using the Chromopainter-Finestructure pipeline to see if it is really "fluked" as you define it. It is currently based on 289k SNP's but I have done experiments that make it possible to do even more subtle divisions with a higher number of SNP's.
DeleteHere is the local Fennoscandian BGA analysis where it is very easy to spot individuals with minority ancestries whatever stated etnicities.
http://fennoscandia.blogspot.no/2013/04/updated-local-fennoscandian-analysis.html
Maju wrote:
"You can't put in the same plane a hard science like Archaeology with soft ones like linguistics or population genetics (yes, genetics is a soft science, scoring similar as economics and well under ecology or agricultural science, not to mention space science). It will depend on the particular facts or claims being compared but in general, what Archaeology produces are facts with high value as evidence while statistical results in population genetics or linguistics are way too often just more or less educated hunches."
More SNPs won't greatly increase the quality of your results, instead larger samples, more different samples and different sampling strategies (if only just for contrast) may and probably will enhance the quality of your results. You are basically repeating the same exercise once and again and, sincerely, I don't see any particular merit nor additional information nor contrast exercise in that. If you want to debate all Europe, you must analyze all Europe (and nearby regions), not just that remote "island" of the Far North. And you should use much more balanced samples (otherwise Nordic clusters would phagocyte others' identities).
DeleteWhat you do is looking at a few trees and ignoring the forest in which they exist.
Even I only have like 5 Saami samples it is enought to make inference about minority ancestries in other Fennoscandians.
DeleteI still can't see you are able to contest any of the inferred ancestries in the project. So far your critism has only been theorical withount any foundation in the actual data.
I'm not fighting to persuade you of anything (other than maybe try better sampling strategies, if nothing else just for contrast).
DeleteI really do not care much about what La Braña composite sequence looks like. There's nothing half clear about it: two individual poor quality sequences (maybe totally different from each other?) were merged to create a slightly better quality composite, which looks just weird not matter how you look at it: ADMIXTURE, ChromoPainter or whatever, with African (but NOT NW African) affinities, which now are Asian affinities and then half-Asian, half-African... It is not even clear if these individuals are Epipaleolithic or Neolithic. There's absolutely nothing clear about the La Braña sequence. Sorry.
It's you who seem somehow obsessed about persuading me of your ideas and "results". I'm not happy about anything in all that: it simply does not make any sense, so please, let me remain skeptic. It is my right.
As I understand it Dienekes DYI calculator based analysis is based on less than 1000 SNPs and my analysis is based on 183k pruned down to 69k. However in your opinion Dienekes K12b analysis is not "fluked" even it is working with less than 1k SNP's and lack many more northern populations than both my own and Vadin Verenich analysis(?)
DeleteI do not know if Dienekes' comparisons produce artifacts or not, they probably do in fact. I'm not particularly fond of his "zombies" either (for example he lets Sardinians cope one of them, even if they are a too homozygous pop., instead he systematically rejects the Basque component, when Basques are clearly much less homozygous than Sardinians or Finns).
DeleteAs I said before, I think that this Braña composite sequence is way too mysterious in too many aspects to draw any conclusions.
But whatever the case it is demonstrated that greater number of SNPs is much less important than better samples:
... according to our results a study employing N = 300 subjects and the Affy500K platform offers higher power than a study employs N = 250 subjects and the Ilmn650K platform. This 20% increase in sample size (N = 300 vs. N = 250) provides more power than the 90% increase in the number of SNPs genotyped (286K SNPs vs. 545K SNPs).
I just did a ADMIXTURE run at K=8 using 4k SNP's from over 2200 worldwide samples in my database. It didn't change the main conclusion from the Chromopainter-Finestructure pipeline analysis.
DeleteCan't see any new post in your feed, so I can't comment.
DeleteIn any case I would strongly suggest you not to upload those hyper-high resolution pictures: in cheap computers like mine at least they really clog the memory up to the point that I had to brute-force-restart my PC twice upon linking to your blog. It'd be much better if you uploaded lower resolution images when possible. I have no problems with most blogs and sites but just very few, particularly yours, being clearly the graphic overload the real problem.
I have made a new post with the results. Sorry for those huge images. I do not know any other way to present them.
Deletehttp://fennoscandia.blogspot.no/2013/07/la-brana-and-modern-world-variation-ii.html
Thanks. I'm just wondering how good is K=8 at cross-validation scores, I feel that more homogeneous regional sub-European components should arise before we can evaluate and that probably requires a K=11 or K=13 level of analysis (but whatever the cross-validation score marks as optimal). But what is clear is that:
Delete1. The affinity towards Saami you detect is caused by two factors: (a) similarly high score of the NW European affinity component and (b) similarly low score of the Mediterranean affinity ("Middle East"??? - Sardinia is not in the Middle East!!!) affinity component.
2. The affinity towards Basques and Danes (respectively) is caused by two factors: similarly low score of (a) Siberian affinity and (b) Central Asian affinity.
3. The orange component, so important and mysterious in La Braña, is shared with mostly peoples of around the Black Sea/Volga area (as well as Orcadians): Chuvash, Turks, Mordvins and Romanians.
4. The third component apparent in La Braña is the blue one that also looks from the Black Sea area (Georgians, Lezgins, Turks, etc.)
Taken at face value these results La Braña people appear North European with a strong Pontic slant. But as regional sub-European components such as the well known Iberian, Basque, Eastern Mediterranean ones have not been resolved (nor also the difference between the Central-North and Finnish components either), I still find the results a bit inconclusive. One could ask: sure there is greater Northern European affinity than modern SW European peoples but this may well be caused ONLY by more recent Eastern Mediterranean admixture in these, so the result is not really helpful in discerning the pre-Neolithic affinities of La Braña.
However it does make them look as pre-Neolithic rather than Neolithic, unless they were exceptionally non-admixed Balcanic peoples, something difficult to argue for but what their composition makes them look to me: people from the Pontic area lacking the Sardinian-Mediterranean component.
So there are still many mysteries to unravel in this.
As for the images, these are fine, what really clogs my computer are some (most) of your ChromoPainter heatmap images. These must have brutally high resolutions like 5000x5000 px or something even greater. So when there are several of these my system slows down to a trickle while downloading them, and sometimes it can't cope at all. I'd suggest loading smaller (500x500 or 1000x1000) images and adding a link to the very high resolutions ones below. You can still upload them to Blogger (in a draft post, never published, my avatar is in one of those, for example) or to an image service like Flickr, etc.
DeleteRegarding the images. They are not at Bloggers but directly linked to Google Sites where they appear at full size and can be downloaded and be viewed and zoomed using local viewer and/or editing programs.
DeleteNot entirely. Saami and Finns did seperate in the Chromopainter-Finestructure as also demonstrated by Vadim Verenich using different method. He had the advantage of having large Saami sample so I couldnt replicate his analysis using ADMIXTURE based tools With so few samples but I could replicate it using the more powerfull Chromopainter-Finestructure pipeline.
DeleteI called it Middle-East because its frequencies peaks among the Bedouins. They live in the Middle-East not on Sardinia.
Maju wrote:
"1. The affinity towards Saami you detect is caused by two factors: (a) similarly high score of the NW European affinity component and (b) similarly low score of the Mediterranean affinity ("Middle East"??? - Sardinia is not in the Middle East!!!) affinity component."
There are no Bedouins in the graphs you published, and Sardinians hold the greatest apportion among your published samples, followed by Turks, Georgians, Italians, Basques, Spanish, Bulgarians...
DeleteHow can you be so sure that this component has anything to do with Bedouins and that it actually peaks among them if you did not use them in the run? Notice that components come and go depending on populations sampled, etc. and other components found in other analysis need not need to be directly extrapolable.
"He had the advantage of having large Saami sample"...
How can that be an advantage? It will totally distort the results. Exaggerating a bit for the sake of clarity, it's almost as if you introduce a large sample of clones, which are a component by default, dragging all others to position in relation to them. I guess you could even get several components out of large enough sample of true identical clones.
Or to put another example, let's say I gather my "army" of almost 40 cousins and sample them, then compare with five others from here and two others from there. Don't you think that most of the results will be about slight differences between my relatives, rather than about differences between them and the unrelated minor samples? I am pretty sure that it will work that well. So a large Saami sample is almost for sure not any help but a burden.
"so I couldnt replicate his analysis using ADMIXTURE based tools With so few samples"
Your non-replication with a more correct number of samples actually seems to me the CORRECT result.
"... but I could replicate it using the more powerfull Chromopainter-Finestructure pipeline."
Which I'm being skeptical about.
As the post titles says its about world variation, I believe it was mentioned in the introduction too. So how you can misunderstand this I do not know and I do not know any other way how to infer African or Siberian admixture in Europeans without samples from these populations. I am interested in how the La Braña and European populations differ from each other vs the rest of the world, not about the internal relationship in Arabia or East-Asia, others can look at those questions.
DeleteSo yes again what I call the Middle-East component peaks among the Bedouins. I believe this have been shown in many research papers using the HGDP population panel.
Maju wrote:
"How can you be so sure that this component has anything to do with Bedouins and that it actually peaks among them if you did not use them in the run?"
ADMIXTURE is about allele frequencies and you need some more samples to get right clustering and inferring the ancestral alleles closest to the truth. In Chromopainter-Finestructure its much less important to have large samples but it is helpful to have at least some to get the tree structure and heatmap boxes right. I have shown this numerous times in Chromopainter-Finestructure where it is very easy to separate Saamis and Finns from each other, but it is not so easy in ADMIXTURE with small Saami sample size.
Delete*Maju wrote:*
"How can that be an advantage? It will totally distort the results"
I appreciate your skeptism, but I would even appreciate your skeptism more if you also had practical experience and not only theoretical experience with the tools I am using. Being sceptical just to be skeptical is not useful or helpful if not backed up by actual knowledge.
DeleteMaju wrote:
Which I'm being skeptical about.
I do have practical experience with ADMIXTURE, which is also a long-time tested tool of which we can know from observation (and also "theoretical" readings) how it behaves in real life. Sample quality and size matters a lot and homozygous populations do tend to distort it so much that, for general analysis purposes, they need to be removed or minimized in weight (sample number). That is a fact.
DeleteADMIXTURE is based on STRUCTURE that has been around for about 10 years. Actually for small datasets my gut feeling is that STRUCTURE still outperform ADMIXTURE. These are powerful but the technology move forward. Today we have software like Chromopainter-Finestructure that is able to work with haplotypes rather than single SNP allle frequencies. I
DeleteThis comment has been removed by the author.
ReplyDeleteThis comment has been removed by the author.
ReplyDeleteI have now narrowed it even further, not only to the common Saami-Finn genetic cluster but to the Saami.
ReplyDeletehttp://fennoscandia.blogspot.no/2013/06/la-brana-and-saamis-ii.html
The K12b calculator analysis by Dienekes using sub 1k number of SNPs is by the way "fluked". I have run my La Braña dataset vs the K12b calculator using 2.5k matching SNP's and got the following:
ReplyDeleteNorth_European 77.34
Atlantic_Med 21.19
Southeast_Asian 1.47
Gedrosia 0
Siberian 0
Northwest_African 0
South_Asian 0
East_African 0
Southwest_Asian 0
East_Asian 0
Caucasus 0
Sub_Saharan 0
As we can see the major "North-European" is far higher using 2.5k SNP than Dienekes sub-1k SNP dataset and as we can see from the minority components they have dissappered. This is the effect of small number of SNP as I have also tested with other genome calculators and the results are very variable for the smaller components shifting between East-Asia, African and Native American but very consistent having huge Northern European ancestral component.
However in the 68k based analysis there is no doubt that there is really a African-like ancestral component.
OK, I take notice of your observations. They are potentially interesting but I rather prefer to wait for new aDNA sampling in SW Europe. For example, a rare open-air Magdalenian burial was just discovered in France and I'm hoping (no news yet) that they do test the DNA. Several sample points will be no doubt more informative than just one low quality composite.
DeleteOops, it's not Magdalenian but Epigravettian (hence related to Italy).
Delete