Open access at Science:
Abstract: We sequenced Early Neolithic genomes from the Zagros region of Iran (eastern Fertile Crescent), where some of the earliest evidence for farming is found, and identify a previously uncharacterized population that is neither ancestral to the first European farmers nor has contributed significantly to the ancestry of modern Europeans. These people are estimated to have separated from Early Neolithic farmers in Anatolia some 46-77,000 years ago and show affinities to modern day Pakistani and Afghan populations, but particularly to Iranian Zoroastrians. We conclude that multiple, genetically differentiated hunter-gatherer populations adopted farming in SW-Asia, that components of pre-Neolithic population structure were preserved as farming spread into neighboring regions, and that the Zagros region was the cradle of eastward expansion.
Broushaki et al., Early Neolithic genomes from the eastern Fertile Crescent, Science 14 Jul 2016, DOI: 10.1126/science.aaf7943
See also...
Economic overhaul + population shift in Late Neolithic Iran
Modeling Steppe_EMBA
Yamnaya =/= Eastern Hunter-Gatherers + Iran Chalcolithic
Iron age Iranian(971-832 BCE) looks less steppe shifted than Modern Iranians.
ReplyDeleteInteresting paragraph:
" We therefore suggest that the source of the non-local gene flow in F38 is more likely to be related to the Anatolians than the
“Samarian” (Yamnaya-like) populations. (see Table S18.5 for exceptions of this pattern detected via further D-statistic. A systematic exception was only ound for the Steppe Sintashta population that still shares less drift F38 than more westward populations, see Table 18.6) "
PCA
https://d2ufo47lrtsv5s.cloudfront.net/content/sci/early/2016/07/13/science.aaf7943/F2.large.jpg?width=800&height=600&carousel=1
I talked about this western Levant-related shift here...
ReplyDeletehttp://eurogenes.blogspot.com.au/2016/07/economic-overhaul-population-shift-in.html
The Indo-European steppe shift came later, most likely during the Bronze Age. You can see it in my models.
http://eurogenes.blogspot.com.au/2016/07/qpadm-tour-of-iran_6.html
1) Their discussion of Yamnaya ancestry caught my eye
ReplyDelete"However, our analyses suggest that Neolithic Iranians were unlikely to be the main source of Near Eastern ancestry in the Steppe population (table S20), and that this ancestry in pre-Yamnaya populations originated primarily in the west of SW-Asia."
Seems like they are also backing the case for a population that is intermediate between Anatolia and Iran as the source, as IRC also raised by Davidski. (My bet is still for most Near Eastern ancestry mainly from a single population basically at the same longitude as the CHG samples we have, but a slightly lower latitude, so avoiding the harsher effects of the LGM, topped up by Anatolian and Iranian Neolithic).
2) The method of inferring ancestry using "haplotype sharing patterns amongst samples using a mixture model" rightly got a bit of stick from Davidski before for using a lack of appropriate references causing misleading conclusions (lack of East Asian samples) without specific caveats. But seems like a powerful method for finding the most related donor populations, provided good recipients are present.
ReplyDeleteFrom their Table S25 - I using Busby's data:
English: CEU - 0.5024, Ireland - 0.3034, Welsh - 0.1942
Polish: Belorussian - 0.1875, Lithuanian - 0.1859, Ukrainian - 0.6266
Basque: Spanish - 1
From their Table S24 - I using Lazaridis's data:
Czech: Hungarian_Metspalu 0.3577, Belarusian 0.2378, English_Kent_GBR 0.2194, Ukrainian_East 0.1479, Croatian 0.0352, French 0.0012, Lithuanian 0.0005, Italian_Bergamo 0.0002
Scottish_Argyll_Bute_GBR: English_Kent_GBR 0.5872, English_Cornwall_GBR 0.3009, Icelandic 0.0642, Norwegian 0.0472, Spanish_Cataluna_IBS 0.0003, French 0.0001
African_American Denver: Yoruba 0.4375, Mende_Sierra_Leone_MSL 0.3422, English_Kent_GBR 0.128, Esan_Nigeria_ESN 0.0912, English_Cornwall_GBR 0.0008, Croatian 0.0001, Icelandic 0.0001, Spanish_Cataluna_IBS 0.0001
Cambodian: Kinh_Vietnam_KHV 0.8457, Bengali_Bangladesh_BEB 0.1276, Thai 0.0157, Dai 0.0087, Kharia 0.0023
With that in mind, I thought it as kind of interesting that when modelling the ancients with recent populations, they got the following:
Loschbour (Busby): CEU 0.3705, Ireland 0.3309, Finnish 0.2467, Basque 0.0518
Loschbour (Lazaridis): Estonian 0.2688, French_South 0.2076, Norwegian 0.1568, Icelandic 0.1445, Lithuanian 0.1201, Scottish_Argyll_Bute_GBR 0.1017, Sardinian 0.0004
KK1 (Busby): Abhkasian 0.3469, Armenian 0.2192, Armenian_LebArmenian 0.154, Georgian 0.1255, Kumyk 0.0961, Pathan 0.053, Meghawal 0.0044, Balkar 0.0007, Armenian_Yegvard 0.0001
KK1 (Lazaridis): Georgian_Megrels 0.3127, Abkhasian 0.312, Kumyk 0.1839, Iran_Fars 0.1087, Pathan 0.0827
(KK1 = Kotias Hunter Gatherer)
WC1 (Busby): Iranian 0.2941, Armenian_LebArmenian 0.2794, Pathan 0.1992, Meghawal 0.1395, Makrani 0.0877
WC1 (Lazaridis): Iranian 0.325, Iran_Fars 0.2246, Sindhi 0.1399, Balochi 0.125, Pathan 0.1174, Lebanese 0.0381, Armenian 0.03
(WC1 = Iranian Early Neolithic)
Ust Ishim (Busby): Bhunjia 0.3818, Egyptian 0.1378, Mawasi 0.1305, French 0.0821, Japanese 0.0751, Kyrgyz 0.0666, Armenian_Yerevan 0.061, Papuan 0.0385, Melanesian 0.0266
Ust Ishim (Lazaridis): Bengali_Bangladesh_BEB 0.2817, Cambodian 0.1552, Egyptian_Comas 0.0972, Mala 0.083, Han 0.0579, Kyrgyz 0.0485, Lebanese 0.045, Tu 0.0363, Italian_Bergamo 0.0332, Italian_Tuscan 0.0262, Papuan 0.0256, Bougainville 0.0241, Spanish_Canarias_IBS 0.024, Jordanian 0.0186, Australian_ECCAC 0.0168, Mayan 0.01, GujaratiB_GIH 0.0095, Onge 0.0048, Mongola 0.0024
(Ust Ishim comes out a real complex stew. Largest components are South Asians with strong ASI, but is this just because they are very admixed in modern day?),
LBK (Busby): TSI (Tuscan) 0.2536, Greek 0.2468, ItalianN 0.1732, Italians 0.1485, Spanish 0.0919, Sardinian 0.0449, SicilianW 0.0307, Tuscan 0.0095, Eastsicilian 0.0008
LBK (Lazaridis): Greek_Coriell 0.2737, Italian_Tuscan 0.266, Spanish_Galicia_IBS 0.2114, Spanish_Baleares_IBS 0.1265, Sardinian 0.0624, Albanian 0.0599
(There is a preference for the LBK sample to prefer Greeks and Tuscan and Spanish for haplotype donation, even though Sardinian is in theory more similar in its unlinked SNP genotypes. I suppose this is the difference between an unadmixed ancestor of modern populations, vs populations that are similar, but not actually as descended from quite the same ancestor?)
Don't you think that it is a little believable that Eastern populations derive from Zagros whereas they peak in the Middle (Indus Valley) and that it is less believable that European populations derive from Barcin (Northern Anatolia) when they peak in the Middle (Sardinia/Italy)? Ridiculous the presupposition that European populations derive from Western Middle East, where no one R1b1 of any kind has been found so far.
ReplyDeleteThe old levantinist prejudice?
Another interesting thing with their Table S24 and S25 and Loschbour is, the analysis part II - that combines donation from moderns and ancients. So you can tell which modern populations can't be modeled well, in haplotypes, as combinations of the moderns in their panel, without additional ancient contribution.
ReplyDeleteSo for those which need additional Loschbour contribution only include:
Lazaridis dataset:
Scottish_Argyll_Bute_GBR: English_Kent_GBR 0.4496, English_Cornwall_GBR 0.3326, Loschbour 0.2177
Icelandic: Norwegian 0.6097, English_Kent_GBR 0.2524, Loschbour 0.0962, French 0.0252, English_Cornwall_GBR 0.0164
Norwegian: English_Kent_GBR 0.3657, English_Cornwall_GBR 0.2454, Icelandic 0.1569, Croatian 0.1378, Loschbour 0.0943
Orcadian: English_Cornwall_GBR 0.527, Norwegian 0.2996, Loschbour 0.0889, Scottish_Argyll_Bute_GBR 0.0846
English_Cornwall_GBR: French 0.5446, English_Kent_GBR 0.4393, Norwegian 0.0123, Loschbour 0.0038
Busby dataset:
Norwegian: CEU 0.4518, Ireland 0.2287, Finnish 0.1817, Scottish 0.098, English 0.0219, Loschbour 0.012, German 0.0059
So, this isn't telling us that these populations have the most WHG related ancestry. We know that is rather the case in Northeast Europe instead.
What it seems to be saying instead is that there are more populations in Lazaridis's dataset that can't quite be covered in their haplotypes by the modern populations in Lazaridis's dataset alone, basically because they need other populations which are mostly like English but with more WHG ancestry to cover their haplotype range. Northeast Europeans, despite greater HG ancestry, don't seem as optimal for this, as a combination of English + Loschbour. Busby's dataset seems to have populations that are like that - Wales, Ireland, Scotland, CEU, so there isn't the same "problem".
This is probably something that can be looked at with the other ancients as well - I see that KK1 seems makes some donation to some West Asian Caucasian populations in the Lazardis dataset in analysis II (looks like this is most the case for Turkish Trabzon, Georgian Megrels, Armenian, Abkhazian), so again, haplotypes might not be able to be quite covered by other modern populations in the dataset.
(Also interestingly, in the analysis with choice of donation from moderns or ancients, Sardinians seem to choose exclusively donation from ancients, mostly from NE1, even though these populations don't choose donation from them...).
@Gioiello
ReplyDeleteThe classic Neolithic farming package came from the Near East, not from Europe or South Asia. It spread with migrants at the expense of local hunter-gatherers.
The Near East later experienced large scale population shifts. And it's not like any of the Near Eastern farmers have been show to carry R1b-M269 or R1a-M417.
So these results showing peaks of early western and eastern farmer ancestry far from their original homelands make sense.
@Davidski
ReplyDeleteReally great to see.
1) Baseded on this, would you be able to see how ASI the neolithic samples are?
2) Would you be able to use Neolithic Iranians as an admixture component for South Asians (including South Indians) and other Western or Central Asians? Would be great to see of Gedmatch as well.
@Davidski,
ReplyDeleteThe Iron Age sample from Hasanlu is likely Urartian (Hurrian-like language) and not likely Indo-European. How do you explain that he is derived for L23 (Z2103)? Autosomally also, he is not Yamnaya admixed.
@aniasi
ReplyDeleteI'm working on a new test for GEDmatch using the ancient Iranian genomes. I'm trying to keep things simple so that it doesn't take too long.
@Roy
You might be right about the Hurrian-like language. But on the PCA the Iron Age Iranian clusters at the edge of the Iranian cluster, far above the Neolithic Iranians, indicating that he does have some steppe-derived ancestry.
I don't find the authors' argument persuasive where they say that D-stats don't point to a steppe origin for the Iron Age individual. It's not like he has to be totally from the steppe; even 15% admix from Yamnaya would still be a lot, but it wouldn't show up well in most D-stats.
Tepe Hasanlu (F38) = R1b1a2a2-CTS1078/Z2103 !!!
ReplyDelete@Gioiello
ReplyDeleteLook at the Neaderthal estimate scores for R*/R1a/R1b clade, page 139 table S21 for clues where the R1b/R1a may have diverged-
Malta1[24.305]0.020049 / 0.017006-depending on study
Villabruna[13.98k]-0.016053
Karelia[8.375K]-0.009587
Now look at Caucasus Hunter Gatherer's they don't score as high as modern day, Dai[0.008542] and Han[0.008776].
Satsurblia[13.225]-0.003929
Kotias[9.72]-0.00566
How can we explain the low Neaderthal scores in the ancient CHG samples?
@Gioiello
ReplyDeleteIf I understand it correctly, that graph showing the haplotype sharing is of Bar8 (Anatolia_Neolithic) vs. WC1 (Iran_Neolithic), so the results are not absolute but relative. For example, the populations in the middle are mostly "white", because they share equally high amounts with both (for example Caucasus populations).
The highest absolute values for WC1 are likely in Iran, not in Balochistan. But because Iran also has higher sharing with Bar8 than Balochi, the circle becomes less green.
For Bar8, the absolute values might probably still be in Sardinia, though. Not sure why they didn't show those absolute values to really know.
Just wanted to add- from
ReplyDeleteThe genetic structure of the world's first farmers-
"We show that the earliest populations of the Near East ....... had little if any Neanderthal admixture..."
@a
DeleteI kind of get lost in some of this, but could "Basal Eurasian" be nothing more than a hole left by varying lack of Neanderthal admixture, or is BE an actual thing?
One thing that I find surprising are the stats with Mota. Have we seen this before? I thought that Mota was an example of unadmixed African (i.e, without Eurasian admixture). But from Table S16.6:
ReplyDeleteD(WC1, EHG; Mota, Chimp) D=0.0789 Z=13.644
D(WC1, WHG: Mota, Chimp) D=0.052 Z=9.339
Is this interpreted as Basal Eurasian admixture in Mota or as Basal Eurasian being closer to Africans (Mota, at least)?
Interesting that we find R1b in Kura Araxes and R1b again in Uratu which directly proceeds Kura Araxes in the same geographical area.
ReplyDeleteI have always suspected that there's not a genetic marker for languages, as in "Indo-european speakers belong to R1a or b" or its mirror image "If he was R1a or b, he must have spoken Indo-european". That R1b Tepe Hasanlu guy fortifies my suspicions
ReplyDelete@bbb
ReplyDelete" but could "Basal Eurasian" be nothing more than a hole left by varying lack of Neanderthal admixture,..."
IMO- Kostenki14 is close to Yamnaya. Perhaps we may find some more clues in ancient Western samples? As it stands, Caucasus and Middle East are showing little to no Neaderthal. Yet we know modern Europeans, and Bell Beakers have Neaderthal admixture.
Here is a good map for seeing the samples location.
http://umap.openstreetmap.fr/en/map/ancient-human-dna_41837#4/47.43/56.56
Study with Oase 1-
http://www.nature.com/nature/journal/v524/n7564/extref/nature14558-s1.pdf
page 9-
"... Thus, Oase 1 shares more alleles with non-Europeans than with post-agricultural Europeans, opposite to the expectation if there was genetic continuity between Oase 1 and later Europeans..."
Oase1-Fu-39.610-0.078461
"It is possible that with more data from Oase 1, a
signal of genetic continuity with later Europeans could be detected. However,it is interesting that to the limits of our resolution, the data are consistent with Oase 1 deriving from a lineage that went extinct in Europe,contributing little or nothing to subsequent populations (unlike Kostenki 14’s population)"
Kostenki14-Kostenki-RU-UP-Fu-37.470-0.025777
Earliest R1b sample in Europe also has Neaderthal.
Villabruna-Fu-13.980-0.016053
One has to come up with a parsimonious explanation; as to how so many modern day Europeans have Neaderthal if Middle Eastern samples are coming in at 0+/-%, and the Caucasus are coming in just slightly higher?
The odd thing is that it increases going East until Denisovan peaks, where it seems it have been absent in the West given this decay or total absence.
DeleteSo basically if David detects a rumble of something ANE like out East, maybe this means the primordial stew that ANE emerged was in a place like Malaysia and it's that brute element that makes it so distinct.
Alberto,
ReplyDeleteMota is admixed, likely with farmers or a pop like farmers. This was in the revision.
Chad Rohlfsen,
ReplyDeleteCould you point us to this finding in the erratum? Per my understanding, the erratum confirmed that Yoruba and Mbuti etc. weren't admixed relative to Mota, but the findings that the latter lacked Neanderthal and hence contemporary Eurasian ancestry weren't affected. According to the D stats in the recent Lazaridis paper, Sub-Saharan Africans (Mota/Yoruba/Mbuti) are no closer to Natufians than even EHG.
https://dl.dropboxusercontent.com/u/26978112/Erratum%20with%20figures.pdf
I think the Iranian Iron Age genome is going to be Iranian Chalcolithic plus significant Armenian Bronze Age or Iron Age admixture. I doubt it's actually Anatolian. Anatolians didn't exist at that time, obviously. It's probably about significant replacement from South of the Caucasus.
ReplyDeleteNeolithic Iranian dating 7,455-7,082 cal. BC WC1 has same Blue eye HERC2 haplotype as do 100% of SHG and WHG samples. Neolithic Anatolians and one CHG with derived allele in rs12913832 had a differnt haplotype.
ReplyDeleteHasanlu is in Mannae. Indo European names are attested in Mannae.
ReplyDeletehttps://en.m.wikipedia.org/wiki/Mannaeans
It is unlikely that there was any ethnolinguistic unity in Mannea. Like other peoples of the Iranian plateau, the Manneans were subjected to an ever increasing Iranian (i.e. Indo-European) penetration. Boehmer's analysis of several anthroponyms and toponyms needs modification and augmentation. Melikishvili (1949, p. 60) tried to confine the Iranian presence in Mannea to its periphery, pointing out that both Daiukku (cf. Schmitt, 1973) and Bagdatti were active in the periphery of Mannea, but this is imprecise, in view of the fact that the names of two early Mannean rulers, viz. Udaki and Azā, are explicable in Old Iranian terms.
Btw who said that everybody in Urartu was urartean.?
ReplyDeleteNo linguist can say that king Rusa is a urartean name. One can't find any academic paper making such a claim.
The name Rusa btw comes from Balkanes
Rusa - A Minoan name written on an Egyptian tablet in the British Museum.
http://www.peiraeuspubliclibrary.com/names/europa/minoan.html
@Gihanga
ReplyDelete"but the findings that the latter lacked Neanderthal and hence contemporary Eurasian ancestry weren't affected."
When that paper came out, it wasn't yet known that Basal Eurasians lacked Neanderthal admixture. So if Mota had admixture from a Basal Eurasian source, it would be difficult to tell without a reference genome for the source.
Gioiello said...
ReplyDelete"Don't you think that it is a little believable that Eastern populations derive from Zagros whereas they peak in the Middle (Indus Valley) and that it is less believable that European populations derive from Barcin (Northern Anatolia) when they peak in the Middle (Sardinia/Italy)?"
If we assume just those two populations they'd expand inward until they bumped into each other somewhere around east Anatolia while expanding outwards east or west with no constraint.
In that model you could end up with 100% red in the west and 100% blue in the east and 50/50 in the middle east border zone so I think it's simply a logical consequence of having two source regions close together.
(Although there's also the question of which LGM refuge the HGs that turned into farmers originally came from i.e. did the HGs that ended up in the Zagros originally expand out of a NW Indian refuge or did Zagros farmers expand and then survive later population turnovers in a NW India refuge?)
About Hurrians
ReplyDeleteStarostin was a genius when he linked Hurro-Urartean to NE Caucasian. This link will be confirmed by genetics. NE Caucasians are loaded by J2 & J1-Z1842. The same is expected for Hurrians.
A lot to chew on here, as the analysis seems more exhaustive than the new Lazaridis et al. preprint (not surprising, considering the focus here is on the Iranian genomes).
ReplyDeleteFor one, this establishes pretty clearly that the Iranian Neolithic samples are to Central/South Asia what the Anatolian Neolithic samples are to Europe. So, when modelling South Central Asian populations, the Near Eastern reference should be the Neolithic Iranian samples, not Chalcolithic Iranians, Iron Age Armenians, or whatever.
Also, our ideas concerning ASI finally get some support in an academic paper. Ust-Ishim eats up the non-West Eurasian ancestry of South Asians, not Han. I think that's pretty significant.
Weird detail though, populations west of the Indus river ("Khurasani" populations, like Pashtuns) don't receive any Ust-Ishim, while closely related yet "classically Indian" populations like Sindhis do get Ust-Ishim. The analysis obviously isn't definitive, but this result is oddly clean, if that makes sense.
@ Grey
ReplyDelete"(Although there's also the question of which LGM refuge the HGs that turned into farmers originally came from i.e. did the HGs that ended up in the Zagros originally expand out of a NW Indian refuge or did Zagros farmers expand and then survive later population turnovers in a NW India refuge?)"
That is just the question. As you know, my theory of an Italian Refugium presupposed that at least during the Younger Dryas at least R1b1-L389, R-V88, R-M335 were in Italy. So far the oldest R1b1a* has been found in Italy (14000 YBP), and all my theory presupposes that the R-L23 found in Samara (and a late sample also in Iran now) came from Western Europe, and above all that the R-L51 subclades came from Western Europe and not from Samara etc.
Thus I think that the question will be decided from the aDNA and not from autosomal calculations. I invited Lazaridis and all the rest to test aDNA from Italy and from Cardials.
@Sein,
ReplyDelete"So, when modelling South Central Asian populations, the Near Eastern reference should be the Neolithic Iranian samples, not Chalcolithic Iranians, Iron Age Armenians, or whatever.'
Affinity to Iran_Neo can be phacilitated through people who were only part Iran_Neo.
@ Sein
ReplyDeleteDid you mean to their Supp Sect. S9
"Among these modern groups, the majority from India have notable additional contributions from the 45 kya Siberian genome Ust’-Ishim,"
Also Papuans , Melanesians, Onge and even some Amerindians (Karitania)
So perhaps Ust-Ishm wasn't a dead branch, as previously thought, although they warn that it could be masking as a proxy
Not sure about what that Mota thing means, but it doesn't look like Anatolia_Neolithic admix into Mota:
ReplyDeleteWC1 Anatolia_Neolithic Mota Chimp 0.0628 12.748
I think this is something quite new that we didn't see before.
Interestingly:
ReplyDeleteWC1 CHG Mota Chimp -0.0064 -0.988
Karl_K
ReplyDeleteWell of course, there's always that possibility, but it doesn't appear that Mota has any significant "farmer like" ancestry.
Also while this hasn't been discussed in much detail realize that Mota, despite owing up to 35% of its ancestry to a divergent Mbuti-like clade, is still significantly closer to Eurasians in general relative to other Africans. That's interesting.
@Gihanga
ReplyDelete'Well of course, there's always that possibility, but it doesn't appear that Mota has any significant "farmer like" ancestry.'
I didn't intend to imply that thevadmixture was from a farmer group. The Basal Eurasian ancestries in Iranian and Anatolian farmers seems to have been seperated by >30,000 years. There could have also been other 'Basal Eurasian' groups in North or East Africa that were even more divergent.
The name 'Basal Eurasian' is likely a complicated misnomer, as Basal Eurasians (and of course the other Eurasians) almost certainly arose from an African source population. The question of back-migration is mainly a geography question.
@Davidski
ReplyDeleteReally great, thanks!
Can we tell how much ASI is in these Neolithic samples?
Also, can someone explain how Ust-Ishim is related to ASI and other populations?
Thanks!
One thing which is probably incorrect in this otherwise great paper (which is actually not the main topic, but draws on the Aegean study by same team) is :
ReplyDelete"Our data show that the chain of Neolithic migration into Europe does not reach back to the eastern Fertile Crescent, also raising questions about whether intermediate populations in south- eastern and Central Anatolia form part of this expansion"
unless I've misunderstood what they're suggesting, the second part of the sentence is false: The main thrust of "European farmers" came from exactly central - eastern Anatolia, not some imgainary "Agean" population
I'll post some significant scores towards Eurasian for Mota later on. On qpAdm Mota is 9% Eurasian in the best fit. About 3% comes back as Onge, so Iranians are closer to Mota than Levantines, due to being significantly closer to the Onge than Natufians or Levant EN.
ReplyDelete@Rob
ReplyDelete"unless I've misunderstood what they're suggesting, the second part of the sentence is false: The main thrust of "European farmers" came from exactly central - eastern Anatolia, not some imgainary "Agean" population"
I don't think they implied that in this paper ( but perhaps they do believe that ). Theyvare just saying that the first farmers of Europe couldn't have come from a group in 'genetic' contact with these Zagros farmers.
Lazaridis models did show Iranian in Anatolians and I get some in EEF. I'm working on that one too.
ReplyDeleteAnyone knows when and where will the genomes be public?
ReplyDeleteUsing the CHG - Iran N composite, Iran CHL produced much worse fits than Iran N for 2 Iraqi Kurd samples, Kurd C1 and Kurd C3.
ReplyDeleteThe following excellent fits were produced using Lazaridis 2016 outgroups in qpAdm
KURD C1
IRAN NEOL 44%
ANATOLIA NEOL 33%
LEVANT NEOL 4%
EHG. 19%
TAIL PROBABILITY. 97%
KURD C3
IRAN NEOL 42%
ANATOLIA NEOL 37%
LEVANT NEOL 0
EHG. 21%
TAIL PROBABILITY. 94%
@ Rob: "Among these modern groups, the majority from India have notable additional contributions from the 45 kya Siberian genome Ust’-Ishim,"
ReplyDeleteNote that's an analysis without any non-Europeans as donor populations, other than Yoruba and Han, and only West Eurasian ancient groups and Ust Ishim.
In that context, it's not necessarily indicative of any contribution from Ust Ishim, really, more the paucity of other moderns to match them well.
Btw, re: my comments upthread re: the analysis of using moderns+ancients to model haplotypes for recent populations, I noticed there was a nice illustration of this in the paper on page 70 of the supplement: http://i.imgur.com/qR6MmTt.png (blurrier and less labelled than I'd like, though).
Thanks Matt
ReplyDeleteI am disappointed , I wanted that Iron age sample to be R1a-Z94! :( .
ReplyDelete@ Nirjhar007
ReplyDelete"I am disappointed , I wanted that Iron age sample to be R1a-Z94! :( . "
Very likely nothing would have changed: always from South Russia ...
@Matt
ReplyDelete"With that in mind, I thought it as kind of interesting that when modelling the ancients with recent populations, they got the following:"
Shame they didn't do it with samples like Andronovo, Sintashta, Srubna, Mezhovskaya, Davidski would probably lose his shit. They'd likely say the same thing Davidski's own calculators say, and that is, these populations were closest to Scandos and NW Europeans rather than R1a Poles/Russians. Eastern Europeans have a lot of unknown HG admixture based on their high Baltic on Eurogenes K15(and I say unknown rather than known WHGs because those don't add up either), unless significant drift happened in such a short period of time which is unlikely.
I suspect this is why he stopped using ADMIXTURE for the less accurate d-stats/TreeMix(lol at Natufians having no SSA affinity/admixture, based on their skulls)
@Matt
ReplyDelete"1) Their discussion of Yamnaya ancestry caught my eye
"However, our analyses suggest that Neolithic Iranians were unlikely to be the main source of Near Eastern ancestry in the Steppe population (table S20), and that this ancestry in pre-Yamnaya populations originated primarily in the west of SW-Asia."
That was obvious a long time ago based o ADMIXTURE runs. Yamnaya had no East Med, CHG had no East Med, Neolithic Iran had significant East Med.
On an unrelated note, one has to wonder who the Sumerians were closest to based on this, Neolithic Levantines/modern Levantines or neolithic Iranians/modern Iranians/Afghans/etc. Lots of bragging rights there.
Shame they didn't do it with samples like Andronovo, Sintashta, Srubna, Mezhovskaya, Davidski would probably lose his shit.
ReplyDeleteYes, TinFoilHatGunther, it's obviously a conspiracy against the superior power of ADMIXTURE and your awesome insights into ancient population structure.
Just in case you're too stupid to get what I just said, I was being sarcastic.
@ Rob, no problem.
ReplyDeleteActually the paragraph goes on to "Several native groups from South
America (Colombian, Karitiana, Pima, Surui) also receive notable contributions from Ust'-Ishim.
However, many of these groups appear genetically differentiated from the prediction based on the inference from analysis (III) (i.e. their levels of “self-copying” are highas depicted by smaller pies in Fig. S29), perhaps suggesting either a substantial amount of population-specific drift in these groups and/or that none of the ancient genomes and modern groups included as surrogates in analysis (III) reflect well the DNA of these groups."
It is still an interesting point for you to raise though, and maybe not as downplayed as I responded above, as it does surprise me that when you have say the following:
Karitiana: Ust Ishim 0.3645, Han 0.6081, "Self Copy" 0.0274
Nihali: Ust Ishim 0.7379, WC1 0.2493, "Self Copy" 0.0128
Melanesian: Ust Ishim 0.5277, Han 0.078, "Self Copy" 0.3943
Colombian: Ust Ishim 0.4703, Han 0.3706, "Self Copy" 0.1591
Onge: Ust Ishim 0.1612, Han 0.0114, "Self Copy" 0.8274
Atayal_Coriell: Ust Ishim 0.0001, Han 0.4529, "Self Copy" 0.547
because I would have expected that even for Karitiana it would've picked up Loschbour haplotypes rather than Ust Ishim, and for South Asia and Onge, much more Han rather than Ust Ishim and "self copy". But it seems like from the haplotype PoV those are too divergent from Ancient North Eurasian and the ENA related contributors to Onge / South Asian to get Loschbour / Han, and the more "basal" Ust Ishim is preferred or the populations are too inbred / bottlenecked / inclusive of close relatives so get high self copy.
@ Gunther: Shame they didn't do it with samples like Andronovo, Sintashta, Srubna, Mezhovskaya
ReplyDeleteNot enough high coverage for those samples I think, as to why that didn't happen. High coverage steppe and EHG samples (preferably also SHG) can't come quick enough, as they feel like the biggest missing piece when looking at these haplotype analyses. I guess there must be technical roadblocks.
I don't know about unknown HG in Eastern Europe specifically (though I think low level contributors from more EHG shifted populations on the Bichon/Loschbour->EHG cline quite likely). It was mainly interesting to me that the haplotypes between the moderns are apparently divergent enough that given the choice of moderns their analysis would prefer to model high latitude NW Europe (Ireland, Scotland, Iceland) with the ancient HG plus a nearby population if need be than just recruiting more HG rich populations for NE Europe.
Related, there is a curious phenomenon in their analysis S9 III for both the Busby and Lazaridis data, where there seems like quite a split between the British Isles and the rest of Northern Europe (both Slavic and Germanic), where the BI seems to pick up quite a bit more of the LBK haplotypes and a bit less of the Loschbour and Kotias haplotypes -
Busby: http://i.imgur.com/tOxZJP2.png (matching neighbour joining clusters http://i.imgur.com/bt2ZMSv.png, note Orcadian gets an odd position due to high "self copy", poss due to relatively high inbreeding coefficient in that pop).
Lazaridis: http://i.imgur.com/kbJdcgo.png (neighbour joining: http://i.imgur.com/0IXbRWA.png)
Particularly kind of interesting because seems true also of the BI populations that were found as the main contributors of haplotypes to Loschbour in their analysis S9 I : Ireland, CEU. LBK haplotypes in this analysis seem to link up pretty the Atlantic region Western Europe (France, Spain, British Isles) for some reason. Even exclusive of even Central and Southeast Europe (Bulgarian, Czech), which seem to pick up NE1 Hungarian farmer instead. Though these aren't literally ancestry proportions.
Compared to there seeming like less of a genetic co-clustering link that can really be found between regional European populations, when we're using these very deep rooted WHG, ANE / EHG, EEF type unlinked clusters - in clustering via levels of those ancient components England doesn't link to Scotland necessarily when I've tried because of different levels of WHG, Poland / Lithuanians not to Czech Republic / Ukrainians necessarily, etc.
@Chad
ReplyDeleteThanks for looking into it. But can you even reproduce those stats with Mota being so significantly closer to CHG/Iran_N than to Anatolia_Neolithic?
These stats are run by the same team that run others previously that were not consistent with what others were getting. Davidski wrote about it here:
http://eurogenes.blogspot.com/2016/06/the-discrepancy.html
If this is another discrepancy in the stats I wonder what's going on there. It's not good for anyone if we're getting different results when using the same samples and tools.
Clustering using the tables S24 and S24 and Section 3 and adjusting out the "Self Copy" fraction:
ReplyDeleteS25 (Busby) : http://i.imgur.com/2ZCYWxd.png
S24 (Lazaridis) : http://i.imgur.com/Mc79PwD.png
The populations from S24 seem a bit more effective at recreating an orthodox tree with relatively fewer unexpected deviations. Possibly because the ratios of West Eurasian admixed Africans to unadmixed Africans to South Asian and Oceanian populations to Native Americans are way different between the two sets, and S25 has more populations overall, even if it is lacking some S24 has.
Just merging those two sets of data from the last post:
ReplyDeletehttp://i.imgur.com/4ncm7NC.png
(Lots of curious features still:
- Caucasians cluster outside the main West Eurasian cluster: This is because they have lots of KK1 haplotypes, while other populations are much more minimal.
- South Asians join in a clade with West Asians and ME: This is because both have WC1 (Iranian Neolithic) haplotypes.
- Papuan-Melanesians join in a clade with some South Asians within the main West Eurasian cluster: That's the Ust Ishim proxy factor pulling them together.
- Many recent Native Americans join with recent Siberians: No specific Native American haplotype cluster
The ability of this one haplotype analysis to conform to classical world variation would be much improved by adding: A South Asian contributor (probably Kharia), a Native American contributor (Karitiana), and a Papuan contributor).
Just in case you're too stupid to get what I just said, I was being sarcastic.
ReplyDeleteouch, thats a true insult. Ive never heard david explicitly say he was being sarcastic.
Rob
ReplyDelete[paper]"Our data show that the chain of Neolithic migration into Europe does not reach back to the eastern Fertile Crescent, also raising questions about whether intermediate populations in south- eastern and Central Anatolia form part of this expansion"
"unless I've misunderstood what they're suggesting, the second part of the sentence is false: The main thrust of "European farmers" came from exactly central - eastern Anatolia, not some imgainary "Agean" population"
To me it reads that because central and eastern Anatolia are mixed they weren't the source into Europe - that came from the unmixed western portion.
This makes perfect sense if there are two (or more) farmer sources close together. Where they expand *towards* each other they meet and unless one has an advantage they stop and form a border and eventually form an admixture along the border. Where they expand *away* from each other this doesn't happen.
.
"not some imgainary "Agean" population"
*If* the now sunken lands were the primary LGM refuges then the Aegean seems like a plausible candidate. They wouldn't be the farmers themselves - just the source of the HGs that became a particular farmer group.
I don't know if it would be helpful or not but one way to define the farmer groups might be
by their HG source region and farmer source region e.g
(just for example, not claiming they're correct)
Aegean / Barcin
Persian Gulf / Zagros
Nile Delta / Natufian
etc
Grey
ReplyDeleteNope
The Anatolian farmers of Europe came from Central & eastern Anatolia predominantly
There are no "sunken islands" ; as the Aegean islands were their present level by the late glacial, thus already by the Neolithic
So we're not missing anything, just a historically incorrect or poorly worded paragraph
@Matt
ReplyDelete"Just merging those two sets of data from the last post:
http://i.imgur.com/4ncm7NC.png"
Splendid tree! The data really support the Elamo-Dravidian hypothesis, particularly if Brahui form an early branch of Dravidian/Elamite (Southworth, "Rice in Dravidian", (Rice, 4: 142-148, 2011). WC1 is in the Elamite speaking area and the linguistic theories suggest that Dravidian might have moved into South Asia/Indus Valley during the Chalcolithic circa 4000-3000 BCE.
That paper by Southworth is open access, and it claims that the Brahui language is an offshoot of proto-Elamitic from the Zagros Mountains. See Figure 2 here.
ReplyDeletehttp://link.springer.com/article/10.1007/s12284-011-9076-9
ReplyDeleteSo no surprise. The sample F38 was the ancestor of YF04770:
YF04770 R-Y16852*
12 24 14 10 11 13 12 12 12 14 13 31 17 9 9 11 11 xx 15 20 28 15 15 16 17 11 11 xx xx 16 15 17 18 xx xx 14 12 11 9 15 16 8 10 10 8 11 11 12 xx xx 16 10 12 xx 16 8 11 22 20 12 xx xx xx x x11 12 xx
Sample ID
HG
6740428
24376846
YF04770
R-Y16852*
G
C
R-Y16852Y19434 * FGC14617 * FGC14610+2 SNPsformed 4800 ybp, TMRCA 4000 ybp info
R-Y16852*
id:YF04770
R-Y11410 FGC14599 * FGC14601 * FGC14608+22 SNPs formed 4000 ybp, TMRCA 1100 ybp info
R-Y11410*
id:YF04142
R-FGC14600FGC14614 * FGC14611 * FGC14600+1 SNPsformed 1100 ybp, TMRCA 600 ybp info
R-FGC14600*
id:YF06610 new
id:YF03999
id:YF03792 ROU [RO-BT]
R-Y21258Y21258/YFS515862 formed 600 ybp, TMRCA 225 ybp info
id:YF06539 new
id:YF05702
id:YF04151
Y-SNP calls from ancient Iran
ReplyDeletePosted on July 17, 2016 by Genetiker
In the table below are links to Y-SNP calls for ancient DNA samples from the Zagros Mountains of Iran.
Sample Period Date BC Haplogroup
AH2 Early Neolithic 8205–7756 J2b-M12* calls
F38 Iron Age 971–832 R1b1a1a2a2a-Y:24376846 calls
I think that very likely he was an Indo-European Iranian person from South Russia.
This is his haplotype:
YF04770 R-Y16852*
12 24 14 10 11 13 12 12 12 14 13 31 17 9 9 11 11 x 15 20 28 15 15 16 17 11 11 x x 16 15 17 18 x x 14 12 11 9 15 16 8 10 10 8 11 11 12 x x 16 10 12 x 16 8 11 22 20 12 x x x x 11 12 x
R1b-M343-L754-L388-P297-M269-L23-CTS1078-L277-Y:22444952-BY3294-FGC32268-YSC0000108-
R1b-M343-L754-L388-P297-M269-L23-CTS1078-L277-Y4366-BY3293-BY3293
R1b-M343-L754-L388-P297-M269-L23-CTS1078-L584-Y:4708905
R1b-M343-L754-L388-P297-M269-L23-CTS1078-L584-PF7580-PF7580
R1b-M343-L754-L388-P297-M269-L23-CTS1078-L584-PF7580-FGC14592
R1b-M343-L754-L388-P297-M269-L23-CTS1078-L584-PF7580-FGC14593/SK2093
R1b-M343-L754-L388-P297-M269-L23-CTS1078-L584-PF7580-FGC14594
R1b-M343-L754-L388-P297-M269-L23-CTS1078-L584-PF7580-S4813-S4813
R1b-M343-L754-L388-P297-M269-L23-CTS1078-L584-PF7580-L943-CTS1848-FGC36759-Y:18080431
R1b-M343-L754-L388-P297-M269-L23-CTS1078-L584-PF7580-FGC14598-FGC14606
R1b-M343-L754-L388-P297-M269-L23-CTS1078-L584-PF7580-FGC14598-FGC14610
R1b-M343-L754-L388-P297-M269-L23-CTS1078-L584-PF7580-FGC14598-FGC14617
R1b-M343-L754-L388-P297-M269-L23-CTS1078-L584-PF7580-FGC14598-Y16852?
R1b-M343-L754-L388-P297-M269-L23-CTS1078-L584-PF7580-FGC14598-Y19434?
R1b-M343-L754-L388-P297-M269-L23-CTS1078-L584-PF7580-FGC14598-Y:24376846-Y:24376846
Closest haplotypes:
ReplyDelete370505 Molchanov Russian Federation R-FGC14598
12 24 14 10 11-13 12 12 12 14 13 31 17 9-9 11 11 25 15 20 28 15-15-16-17 11 11 19-23 16 15 17 18 38-38 14 12
and I wrote a lot about these Ashkenazi Jews as Goldschlager:
298652 Juda Spira, b. ~1750 and d. 1820, Trebic,Czech Rep R-FGC14618 12 24 14 11 11-13 12 12 13 13 13 29 16 9-10 11 11 25 15 20 28 15-15-16-19 12 11 19-19 16 15 17 17 36-37 12 12 12 9 15-16 8 10 10 8 10 11 12 23-23 16 10 12 12 18 8 12 22 20 15 12 11 14 11 11 12 12
271988 Shimshon Goldschlager b ~1800 d1879 Botosani (Y) R-FGC14600
12 24 14 11 11-13 12 12 13 13 13 29 16 9-10 11 11 25 15 20 28 15-15-16-19 12 11 19-19 16 15 17 17 37-38 12 12 12 9 15-16 8 10 10 8 10 11 12 23-23 16 10 12 12 18 8 12 22 20 15 12 11 14 11 11 12 12 36 15 9 16 11 26 26 17 12 11 12 12 10 9 13 12 10 11 11 30 12 13 24 13 10 10 21 15 19 14 23 19 13 17 24 12 23 18 10 14 17 8 11 12
Elamite-Dravidian theory is not scientific .
ReplyDeleteBTW , Its indeed significant that Zoroastrians show great affinity .
Gioiello ,
I think indications are going the opposite of your suggestion . Think deeply.
@ Nirjhar0007
ReplyDelete"Gioiello ,
I think indications are going the opposite of your suggestion. Think deeply".
What are you saying? About what?
Have you read my posts above? P38 is an R-L584-PF7580 with a private SNP in common with a Noble Russian of to-day (thus it belonged to an Indo-European Iranian come from South Russia). The Jewish Ashkenazi R-L584 of Goldschlager has a MRCA only 1100 years ago, thus introgressed or from Khazars or from some other... we'll see.
Are you referring to the migration from Iran to India? Linguists know from so long that Dravidian languages came from West (Brahui is a relict) and were linked to Elamite. I know that from Alfredo Trombetti, Elementi di glottologia, 1923.
I was referring to your South of Russia suggestion .
ReplyDeleteBTW I am sorry but Brahui is not a ''relict'' , it came there in middle ages . Other Dravidian groups even reached Nepal! .
Oh yes . Genetikers ''calls'' are not very trust worthy , I trust this person more -
ReplyDeletehttp://www.ancestraljourneys.org/wasianneolithicdna.shtml
That person (do you know which is her Italian nickname?)catches from Others what she likes, and you do continue not believing Genetiker and smal (Sergey Malyshev) not believing him about Villabruna... I have explained above how matters go, not only, but it is all in line with my theories, also the introgressed Jewish Ashkenazi haplotypes. Have you seen how many samples of Y are tested out of Italy? In Italy only 5 so far...
ReplyDelete@ Nirjhar007
ReplyDeleteIf you say that about Brahui and Dravidian languages, perhaps you know some Indian dialect, but nothing about Brahui or Malayalam...
Your theory isn't the only one:
ReplyDeleteThere is no consensus as to whether Brahui is a relatively recent language introduced into Balochistan or remnant of an older widespread Dravidian language family. According to Josef Elfenbein (1989), the most common theory is that the Brahui were part of a Dravidian invasion of north-western India in 3rd millennium BC, but unlike other Dravidians who migrated to the south, they remained in Sarawan and Jahlawan since before 2000 BC. However, some other scholars see it as a recent migrant language to its present region. They postulate, that Brahui could only have migrated to Balochistan from central India after 1000 CE. The absence of any older Iranian (Avestan) loanwords in Brahui supports this hypothesis. The main Iranian contributor to Brahui vocabulary, Balochi, is a Northwestern Iranian language, and moved to the area from the west only around 1000 CE. One scholar places the migration аs late as the 13th or 14th century.
and it could be as your OIT (ahahahahahahah!)
That person is more accurate in my observation . Just not Italy , there are many areas which are under sampled . We still have miles to go, before we conclude that, we have reached enough coverage ,to confidently infer on Eurasian prehistorical matters ! .
ReplyDeleteAnd we have to put linguistics out of it . For linguistic matters we need a combination of Archaeological+Anthropological( including of course aDNA) +Textual collaboration.
But we have to be careful . There many biased and ''well sold '' hypotheses , meaning they if we observe closely, they are not very scientific! .
At the moment via aDNA we are only judging on population migrations and replacements and admixture , but its still in very early stages .
At the moment apart from the Indian aDNA , I am looking forward to the West European and South European genomes.
This paper that I linked to before about the word rice in Dravidian is remarkable in the way that it matches the ancient genomics results that we've seen from Iran.
ReplyDeletehttp://link.springer.com/article/10.1007/s12284-011-9076-9
Clearly, Brahui and surrounding South Asian groups, as well as Dravidians from South India, have a lot of ancestry from the proto-Elamite homeland in Neolithic western Iran.
It could all be one big fat coincidence, but I don't believe in big fat coincidences.
Yes, My theory is not the only but the better one . Because its not based on an theory itself Dottore, but on linguistic observations!.
ReplyDeleteI don't support OIT , how many times I have told you this ? :) .
But that Brahui came from India and not the other way around is an OIT anyway, and it is false.
ReplyDelete@Arame
ReplyDeleteThere are Indo European names in Manneae yes but don't jump on wrong conclusions. The sample predates the Indo European influx into Mannea by centuries. The Indo European influx is Scythian/Cimmerian. Now I don't think those are known to be R1b z2013.
Dottore,
ReplyDeleteIf something like Linguistics has any value . Its the truth . Even staunch AIT supporters like Witzel agree with it .
BTW Why every one is underestimating that Zoroastrians show great affinity?.Why it is not considered that they propose a connection of Zagros neolithic with Indo-Iranians?.
@ Nirjhar007
ReplyDelete"Why every one is underestimating that Zoroastrians show great affinity?.Why it is not considered that they propose a connection of Zagros neolithic with Indo-Iranians?".
Because genetics is against your theory. We have:
1) R-L23-Z2105 in Samara from 3339-2917 BC [Yamnaya Russia Lopatino I, Sok River, Samara [I0429/SVP 38] M 3339-2917 BC R1b1a2a2* (Z2103) Z2105+, L23+, L150+, M269+, L584- T2c1a2
Haak 2015; Mathieson 2015; Sergey Malyshev]
2) P38 (971-832 BC) in Zagros with a private mutation (Y:24376846 xxx) found also in Russian Molchanov (FTDNA: 370505) but not in other subclades like Francisco Grijalba (FTDNA: 351640) very likely linked to the introgressed Jewish cluster of Goldschlager (FTDNA: 271988) with a MRCA 1100 years ago. Grijalba may have come from Visigoths of Iberia come from Eastern Europe.
Of course Molchanov may have come from Scythians of South Russia, but the oldest R-L23-Z2105 has been found in Samara and not in Iran or elsewhere in Middle East so far, thus the origin from South Russia is the most likable.
We are waiting fo other data which may change my opinion.
No its not against my theory/observations . Yes. to change your mind we wait for some demonstrations from Y SNP side .
ReplyDelete@ Nirjhar007
ReplyDeleteI appreciate your honesty, however, I don't think you were around when Gioiello was banned when he spoke his mind.
I can't remember if it was around the time of this gem.
anyway you can look for yourself what the study has to say
figure 1- notice the tree with R1b and R1a/ in light of the new data about Iron Age R1b sample[Iron age Iranian(971-832 BCE) looks less Steppe shifted than Modern Iranians.]
http://www.nature.com/ncomms/2013/131217/ncomms3928/full/ncomms3928.html
From the haplotype sharing tables (S24 and S25) that Matt has already commented extensively above, another interesting thing is to check SSA admixture in Eurasians using this other approach. From Analysis III (only ancients + Yotuba and Han as donors) from the Lazaridis set, the Eurasians + North Africans that get Mota admixture (+ Yoruba in brackets if it's the case):
ReplyDeleteYemen: 18.2% (+3.4% Yoruba)
Saharawi: 15.5% (+8.1%)
Egyptian_Metspalu: 15.2% (+4.5%)
Egyptian_Comas: 14.1% (+3.8%)
Tunisian: 13.9% (+7.9%)
BedouinA: 12.7% (+0.2%)
Mozabite: 10.2% (+6.9%)
Algerian: 10% (+6.3%)
Jordanian: 10%
Syrian: 8.8%
Palestinian: 8.2% (*25% "SelfCopy", not sure if/how that affects the numbers)
Saudi: 8%
Lebanese: 6.8%
BedouinB: 6.7% (*40% "SelfCopy")
Maltese: 4.4%
Spanish_Canarias: 4.1% (+1%)
Italian_WestSicilian: 3.1%
Spanish_Extremadura: 3%
Italian_EastSicilian: 2.9%
Druze: 2.9% (*38% "SelfCopy")
Spanish_Murcia: 2.7%
Spanish_Castilla_y_Leon: 2.7%
Spanish_Andalucia: 2.5%
Spanish_Galicia: 2.5%
Cypriot: 2.4%
Spanish_Castilla_la_Mancha: 1.9%
Spanish_Valencia: 1.6%
Spanish_Cantabria: 1.6%
Turkish_Adana: 1%
Sardinian: 0.14%
Makrani: 0 (+2%)
The numbers in the Busby set are pretty much the same, with these exceptions that in Lzaridis get 0% but in Busby get some admixture:
Balochi: 0 (+1.3%)
Brahui: 0 (+1.6%)
Sindhi: 0 (+1.7%)
Iranian: 4.8%
And not present in Laz:
Moroccan: 9.8% (+11%)
UAE: 8.4% (+4.5%)
Overall quite in agreement with what we've seen in Admixture over the years. Some numbers might look a bit higher due to Mota being less divergent than Yoruba or Mbuti used in many admixture clusters.
Elamo-Dravidian is far from being an accepted theory. The main stream view is that the IVC was non Dravidian On the Genetic Affiliation of the Elamite Language - https://www.google.co.in/url?sa=t&source=web&rct=j&url=http://starling.rinet.ru/Texts/elam.pdf&ved=0ahUKEwjYwNH6k_zNAhVY5WMKHdIADckQFgheMBE&usg=AFQjCNHNdjp_w31OUHw6MH8IdG8NBTFZzg&sig2=vUdamhDl3x2bhum_yRmXzg
ReplyDeleteThe majority of loans in the Rg Veda are from some extinct family of prefixing languages not Dravidian, making it unlikely that Dravidian was the language of the IVC. Further Dravidian is usually considered a relatively young family with a time depth of only 4500 - 3500 years. Brahui is considered an emigrant from central India on the basis of the community legends and also the fact that all iranic loans are from balochi a late immigrant to that area with no trace of older east iranic influence
ReplyDeleteThis comment has been removed by the author.
ReplyDeleteKurti
ReplyDeleteBtw have You noted that Armenians are slightly blueish rather than reddish. Remember I was telling You Armenians will have as much strong affinity to Iran Neolithic as they have to EEF/ANF. You were not believing me.;)
"Also interesting that long ROH indicating consanguinity are detected in an ancient population for the first time, probably due to the changed incentives surrounding marriage now that property and inheritance become a major part of human life post-Neolithic."
ReplyDeleteIt's not the first time, they found this in Neanderthals. And it might have a lot to due with the fact that very few ancient genomes have high enough coverage and quality to detect ROH even if they were there.
@Jijansu
ReplyDelete"The majority of loans in the Rg Veda are from some extinct family of prefixing languages not Dravidian, making it unlikely that Dravidian was the language of the IVC."
But you are assuming that the loans in the Rig Veda are from the IVC. You have no way of knowing if this is true. These people passed through a lot of territory in a short amount of time, and the loans could have come from many different places.
@Karl
ReplyDelete"These people passed through a lot of territory in a short amount of time, and the loans could have come from many different places."
Perhaps you should read up mainstream IE linguistics. The rigvedas and the later brahmanas have the least amount of loans of all IE languages. linguists have had to use a magnifying glass. The few loans there are considered indic. This is probably also true of the Avesta as well, probably no loans at all but a much smaller body of work.
Most enemies/neighbors of note, for vedic speakers were also IE for the most part.
In generlal such highly inflecting languages might indicate an inbred linguistic hegemony with very little substratum or adstrate.
@gioello
"P38 is an R-L584-PF7580 with a private SNP in common with a Noble Russian of to-day (thus it belonged to an Indo-European Iranian come from South Russia)."
there is no I-Ir language evidence from south russia. Avestan is a genuinely attested language and Zoroastrians are a genuine zombie/fossil population. Perhaps everything goes for you. You thought Uzbekistan is near the caucasus.
As for elamo-dravidian and brahui etc. Its inconclusive fuzzy crap. Brahui could be from prehistoric residents or recent emigrants from central india. Anything is possible.
@postneo
ReplyDelete"Perhaps you should read up mainstream IE linguistics. The rigvedas and the later brahmanas have the least amount of loans of all IE languages."
I wasn't the one saying there were any loans at all. I was merely mentioning that there is 0% reason to suppose that if any loans actually did exist, that they would have from the IVC.
ReplyDelete@ postno (seen that you write “gioello” for “gioiello” I may write “postno”)
I didn't reply to your provoking words, and thought that you didn't consider me as I have no interest for you, but my last words to you:
1) about R1b1-L278* found in India (I remember to you that was I to test Raza and Joshi) I said that I didn't think that India was at its origin, but that it came very likely from Central Asia, also Uzbekistan when it is found to-day too, and that I thought that its origin is in the Caucasus, where it is found to-day too with other linked haplogroups like R1b1-L389+ (YCAII=21-23) not found so far in India. Your irony has no meaning.
2) P38 from Iron Age in Iran (less than 3000 years ago) is very likely linked to Indo-Iranian speaking people migrated to Iran from South Russia in that period (everything except you know that and that South Russia has many Iranian speaking peoples like Scythians, Cimmerians etc.). That we have a Noble Russian with a private mutation in common with him (the subclade is old 4000 years as to YFull, may demonstate that he may descend from linked people remained in Russia, and that a subclade of that haplogroup which lacks that private mutation and found in an Iberian (Grijalba) may demonstrate an origin from Visigoths who came from eastern Europe (of course the Jewish linked cluster, old only 1100 years, introgressed from there, even though that dislikes “a” and his acolytes).
3) About the origin of Dravidian languages and the link with Elamite, of course the paper of Starostin (Russia has many great linguists, do you know someone from India?), that someone posted above, is right. Elamite is very likely an isolated remnant of a linguistic group intermediate between Nostratic and Afro-Asiatic more than closely linked to Dravidian (which may belong to Nostratic itself). Starostin believes also to an Eurasian family which comprehends Nostratic, Afro-Asiatic and Sino-Tibetan, and won't be I to deny that, being a follower of the monogenesis of the language of Alfredo Trombetti. But that has nothing to do with the fact, posted from Davidski, that there are strong proofs that Brahui didn't come from India.
4) Someone, perhaps a compatriot of yours (at least at the Indo-Iranian level if not to the only Indian one), posted that results from Rakhigarhi are coming out. The bet is in: I opt for C, D, P, Q, O, H, and many extinct haplotypes.
Gioiello,
ReplyDeleteCalm down . No, there is no evidence that they didn't come from C India . They also have relatively small % of Dravidian vocabulary (~15% if i'm not wrong) , which is something we expect from a outlandish group. This is something accepted by linguists and other migrations also reached much Northern territories .
BTW your bet is pathetic ...
Let the aDNA demonstrate .
@RK
ReplyDeleteYes, the stats with Mota are very interesting, but I'd like to know if they are real. When I see some other strange stats, like:
WC1 EHG Anatolia_Neolithic Khomani -0.027 -8.016
WC1 EHG Greek_EN Khomani 0.0325 5.013
It doesn't inspire me a lot of confidence. I don't know what can it be, maybe the way these samples are genotyped? (Greek_EN and WC1 are both from this same team).
@ karl_k
ReplyDeleteThe geography of the rig veda is largely confined to the plains of the Indus and ghaggar-Hakra (saraswati) basins (a few early verses may reflect some memories of Afghanistan) which overlaps with ivc territory. You may read Witzel's extensive treatment of the substrates in Vedic if interested. If the IVC was Dravidian speaking why have they contributed so negligibly to Rg Vedic?
@Gioiello
While genetics can trace population movements it must align itself with linguistic theories while discussing linguistic change. As regards Brahui mainstream opinion suggests a movement from central India. Such a movement need not necessarily have left behind a detectable genetic footprint
@Jijnasu
ReplyDelete"If the IVC was Dravidian speaking why have they contributed so negligibly to Rg Vedic?"
I'm sorry. I didn't know that you do not understand English very well.
I do not think that the IVC spoke Dravidian. I am only saying that it might be an impossible question to answer, due to lack of writings.
karlk
ReplyDelete"I wasn't the one saying there were any loans at all"
OK so you first said the vedics travelled a lot and picked up layers of loans
then you said their loans are not from IVC
mainstream AIT says that vedics came in from IVC areas. perhaps you are a proponent. But then you saying they picked up no loans when they traversed IVC. The paltry loans in vedic loans have been termed para munda/language x, what area did they traverse to pick those up?
@postneo
ReplyDeleteNo problem. I see that you can not follow. Forget about it.
I did not ever mean to imply that anyone ever DID pick up a loan. Only that IF they did, then no one alive today could say from where it had been picked up, and they certainly could not say that it had anything to do with the IVC, as we do not know anything about the language that they spoke.
Something is cooking up in Western Europe , probably will hear soon .
ReplyDelete@Davidski
ReplyDeleteRe: Rice in Dravidian. Interesting article. Although some things not as certain - i.e. pinda which author presumes from Proto-Dravidian.
However, see Armenian pind https://en.wiktionary.org/wiki/%D5%BA%D5%AB%D5%B6%D5%A4
@Nirjhar
As Elamites were non-IE and as we have significant documentation that their language was replaced by Iranian with the Achaemenids, it seems unlikely that Zagros = I-I.
More likely, I-I back to Bactria and further up.
But Zagros is a game-changer. And EMBA Steppe too.
Basically will need to wait for Bactrian and IVC.
Also, the idea that IVC spoke a complete unknown language that left no trace... yeah, no. IVC was a massive civ that dwarfed most B.A. civs.
Yes Atrior, we wait for more data from aDNA side . We badly need something from India and SC Asia now . It is of course not advisable that IVC/SSVC spoke an extinct language and its people had extinct hg's...
ReplyDeleteYes similar thing happened in India where Dravidians of South were seriously effected by the IE's of north . Only they didn't become extinct .
Oh yes, to be more accurate , the IE influence in south was positive.
ReplyDeleteBTW , Atrior, since you have good knowledge in these matters . I have a question for you , I was having this debate with some friends that can Zarathustra be regarded as Persian or SC Asian (Around Bactria) , its a bit difficult to arrive in conclusive opinion, but whats your take?.
@Nirjhar007
ReplyDeleteIt's not that inconclusive at all. He was SC Asian/Aryan/Bactrian. Zoroastrianism moved into Persia, where his doctrines had a much larger influence.
But again, it's dangerous to think in terms of modern connotations of "Persian" or "SC Asian."
Better to think in terms of Airyanem Vaejah in this respect.
As someone said that I don't understand anything about genetics, and as from many years (but later than 1998) I am saying that Jews of to-day have a little to do with old Jews, and as, when a Jew matches an Iberian, Jews say that the Iberian is a Crypto-Jew and not the other way around as I am supporting from so long, I'd want to complete my post about P38:
ReplyDelete1) the oldest haplotype who matches P38 (found in Iron age Iran from Indo-Europeans migrated from South Russia) are Russian Molchanov
370505 Russian Federation R-FGC14598
12 24 14 10 11-13 12 12 12 14 13 31 17 9-9 11 11 25 15 20 28 15-15-16-17 11 11 19-23 16 15 17 18 38-38 14 12
and Sliman (who has nothing to do with Jews)
170600 Unknown Origin R-FGC14598
12 24 14 11 11-14 12 12 13 12 13 29 17 9-10 11 11 24 15 20 28 16-16-16-17 11 11 19-23 16 15 19 17 35-38 13 12 9 9 16-16 8 10 10 8 11 10 12 23-23 15 10 12 12 17 8 12 22 20 13 12 11 13 12 11 12 12 34 15 9 16 12 25 26 17 12 11 11 12 10 9 12 12 10 11 11 30 12 13 24 13 10 10 22 15 18 13 23 19 12 15 21 12 23 18 9 14 17 8 11 11
and is linked to Sahadi from Lebanon who may descend from Crusaders or from Armenians (private SNPs will say)
131176 Najeeb Sahadi b. 1879, Zahlé, Lebanon Lebanon R-FGC14598
12 24 14 11 11-14 12 12 12 12 13 29 16 9-10 11 11 25 14 20 28 15-16-16-17 12 12 19-23 16 15 18 17 37-38 13 12 9 9 16-16 8 10 10 8 12 10 12 23-23 15 10 12 12 16 8 12 22 20 13 12 11 13 12 11 12 12
Iberians like Salido, Grijalba etc
351640 Francisco Grijalba, b. 1801 Unknown Origin R-FGC14595
12 24 14 11 11-13 12 12 13 13 13 29 16 9-10 11 11 25 15 20 28 15-15-16-19 12 11 19-19 16 15 17 17 36-38 12 12 12 9 15-16 8 10 10 8 10 12 12 23-23 17 10 12 12 18 8 12 21 20 14 12 11 14 11 11 12 12 36 15 9 16 11 25 26 17 12 11 13 12 10 9 13 12 10 11 11 29 12 13 24 13 10 11 20 15 19 14 23 18 13 17 25 12 23 18 10 14 17 8 11 12
very likely descended from Visigoths
and someone linked with them is the ancestor of the Jewish clade very recent and with a MRCA as to YFull at 1100 years ago:
298652 Juda Spira, b. ~1750 and d. 1820, Trebic,Czech Rep Czech Republic R-FGC14618
12 24 14 11 11-13 12 12 13 13 13 29 16 9-10 11 11 25 15 20 28 15-15-16-19 12 11 19-19 16 15 17 17 36-37 12 12 12 9 15-16 8 10 10 8 10 11 12 23-23 16 10 12 12 18 8 12 22 20 15 12 11 14 11 11 12 12
187115 Wolf SPIER c 1690 d. Merzhausen Germany R-FGC14600
12 24 14 11 11-13 12 12 13 13 13 29 17 9-10 11 11 25 15 20 28 15-15-16-19 12 11 19-19 16 15 17 16 37-37 12 12
271988 Shimshon Goldschlager b ~1800 d1879 Botosani (Y) Austria R-FGC14600
12 24 14 11 11-13 12 12 13 13 13 29 16 9-10 11 11 25 15 20 28 15-15-16-19 12 11 19-19 16 15 17 17 37-38 12 12 12 9 15-16 8 10 10 8 10 11 12 23-23 16 10 12 12 18 8 12 22 20 15 12 11 14 11 11 12 12 36 15 9 16 11 26 26 17 12 11 12 12 10 9 13 12 10 11 11 30 12 13 24 13 10 10 21 15 19 14 23 19 13 17 24 12 23 18 10 14 17 8 11 12
and so on.
Atriðr ,
ReplyDeleteI like it . You gave a straightforward answer! :) . I agree with you.
Gioiello ,
For gods sake , stop it! . We get it , you have a point regarding SNP's . But the let more aDNA come, so we can get to a finer resolution .
Atriðr ,
ReplyDeleteYou know this research?.
http://booksandjournals.brillonline.com/content/journals/10.1163/24055069-00102003
A new paper on the Y dna N
ReplyDeleteBehind pay wall
http://www.cell.com/ajhg/abstract/S0002-9297(16)30160-4
If any bud is interested -
ReplyDeletehttp://www.nature.com/articles/srep29890
@karlk
ReplyDelete"But you are assuming that the loans in the Rig Veda are from the IVC.
You have no way of knowing if this is true. These people passed through a lot of territory in a short amount of time, and the loans could have come from many different places."
"I did not ever mean to imply that anyone ever DID pick up a loan.
Only that IF they did, then no one alive today could say from where it had been picked up, and they certainly could not say that it had anything to do with the IVC, as we do not know anything about the language that they spoke."
------
I am not saying that its from IVC or hazarding a guess on the language of IVC. but what you are saying is that they picked up some paltry loans from unknown regions but had special loan filters on when they traversed IVC areas. The loans are not elamite, uralic, semitic, dravidian etc. So the best way out for you is to say things like:
its a snapshot of a state before any IVC type loans could occur
it became a priestly language that was conservative about loans during the indic phase ranging btween 600 to a 1000 years
@Gioiello
Cimmerian is not a language. There are personal name from assyrian inscriptions from Iraq are very far from southern russia. There are no I-Ir inscriptions from southern russia. In any case 700 BC age of such people is not old enough for being the ancestor for I-Ir.
@ postneo
ReplyDeleteThis knowledge is common here, it is enough you read Wikipedia:
"The Cimmerians (also Kimmerians, Greek Κιμμέριοι Kimmerioi) are an ancient people, first mentioned in the late 8th century BC in Assyrian records.
Likely originating in the Pontic steppe and invading by means of the Caucasus, they probably assaulted Urartu in c. 714 BC. They were defeated by Assyrian forces under Sargon II in 705 and turned towards Anatolia, conquering Phrygia in 696/5. They reached the height of their power in 652 after taking Sardis, the capital of Lydia. Soon after 619, Alyattes of Lydia defeated them. There are no further mentions of them in historical sources, but it is likely that they settled in Cappadocia.[1]
The origin of the Cimmerians is unclear. They are mostly supposed to have been related to either Iranian or Thracian speaking groups which migrated under pressure of the Scythian expansion of the 9th to 8th century BC.
According to Herodotus, the Cimmerians inhabited the region north of the Caucasus and the Black Sea during the 8th and 7th centuries BC (i.e. what is now Ukraine and Southern Russia), although it isn't possible to identify the Cimmerians as the bearers of any specific archaeological culture in the region".
Anyway it seems that aDNA from India is ready to publish. Thus let's wait and all will be clear.
P.S. I spoke of Iranian languages, not I-Ir. But that is worst for you. Indian languages separated before of course...
@postneo
ReplyDelete"you are saying is that they picked up some paltry loans from unknown regions but had special loan filters on when they traversed IVC areas"
That is not at all what I have ever said. It is as if you don't understand half of the words that people write.
@karl
ReplyDeleteI understand pretty well just Karl, just extrapolating a little, keeping folks on the straight and narrow. According to the mainstream, vedics spent at least 600 years composing verses near IVC regions practicing "elite dominance" and transferring language.
@gioello
please read the wiki entry you posted. they are a unknown people of uncertain language and unknown origin first recorded in assyria, phrygia and Lydia. please look these places up them on a map. The word uncertain is used so many times in all of their supposed activities. They are from 800 BC at most. You are building linguistic theories on rubbish.
Guys
ReplyDeleteF38 autosomaly doesn't look Cimmerian or any recent Steppe invador. Although he has some Steppe affinities.
He is most probably from North East Anatolia.
This Hasanlu F38 can be remnant of IA Mushki expansion. The absence of strong Levantine affinities means he was not from Armenia MLBA and certainly not places close to Levant. Most probably from North East Anatolia.
This is top5 of f3 stat
Georgian
Abkhazian
Armenian
Albanian
Iran_Zoroastrian
But off course if the basal SNPs of L584 are in that Russian noble person then off course it means that he came from Steppe but 1000 years earlier. Not in Iron Age.!
ReplyDelete@ Aram
ReplyDeleteOf course P38, found in that place and in that age, may be thought belonging to Medians, and of course also Medians may have come from South Russia but 1 thousands years before as yous say and have had many vicissitudes we don't know, also having inhabited in the Caucasus for so long...
@ Davidski
ReplyDelete"I honestly couldn't give two figs about R1b.
Tell us what he said about R1a. Oh wait, R1a did spread from the steppes. Haha."
Probably secondly, and linked with Indo-European satem languages, but we couldn't understand why the oldest subclades of R1a-M420* are in Western-Central Europe and Middle East if it spread from the steppes. I am afraid that Italy and Western Europe has something to do with it too... and don't do like Nirjhar007's friends: let's wait for the aDNA, which will decide anyhow.
P.S. I have many fig trees in my garden and it is a pity that they grow only on Mediterranean countries. In my wife's island grow above all Indian figs, but I never ate them.
I guess Y-DNA R1b is a related topic here . So, I just heard from a certain researcher friend of mine , that a certain geneticist from Harvard gave a speech recently and the speaker suggests R1b didn't spread from the Steppes!.
ReplyDeleteI think quite a few here know of what I'm talking about . Sorry can't be more specific :( .
You're just confused as usual.
ReplyDeleteYou're just confused as usual
ReplyDeleteHow? by notifying you guys, that a geneticist recently spoke something, that indicates there are some unpublished samples, which backs his suggestion regarding a certain y-snp's history?.
@Nirjhar,
ReplyDeleteHummm you dont happen to know where those unpublished samples are from? :)
@Gioiello
ReplyDelete"Anyway it seems that aDNA from India is ready to publish. Thus let's wait and all will be clear."
What!? What!? What!? Really!?
I am very happy to hear this. Do you know where this is reported?
@ aniasi
ReplyDeleteIt was announced here:
udaya udaya ranasinghe seneviratne has left a new comment on the post "The genetic structure of the world's first farmers...":
Rakhigarhi DNA clues are out. Oficial results will come in next Nature Magazine. What we have to do now is rehabilitate. It seems no archeogeneticist involved in this thread beforehand.
Hope to see them soon and that they don't end like Tut's haplotype.
Hummm you dont happen to know where those unpublished samples are from?
ReplyDeleteOut of his ass.
Rakhigarhi DNA clues are out. Oficial results will come in next Nature Magazine. What we have to do now is rehabilitate. It seems no archeogeneticist involved in this thread beforehand.
The new Nature magazine came out today. Nothing about India.
http://www.nature.com/nature/current_issue.html
Would be great if it's true that the Rakhigarhi samples are coming out soon. Speculations have been going on for too long about South Asia. Let's get some reality at last.
ReplyDeleteWhat about these Iranian samples? They must have been available for a while if others were testing them. It would be interesting to see that Iron Age sample carrying R1b. Or the high coverage Zagros sample.
Any further information coming through?
ReplyDeleteIs there a reason why Kurdish group are not mentioned at all in this work, regarding similarity or lack thereof with WC1? Lots of specific groups are mentioned on page 501, including not just general "Iran (Fars)" and Afghan, but also specific subgroups like Balochi and Brahui. Given that Kurds are the main Iranic"-speaking (and once formerly Zoroastrian) peoples west of Iran, their lack of mention seems to imply that they are not as close to WC1 as any of the other several populations mentioned. If that implication is untrue, then they should have been mentioned explicitly.
ReplyDeleteRegarding modern populations with affinity to WC1, Yemen is odd in not having any Indo-Iranian (except for one dynasty) nor Zoroastrian connection. Not that such a connection is required, but all other populations have it (e.g. Armenians share a Zoroastrian past, as do Uzbeks, Tajiks, etc.).
I also wonder if the populations said to share >10% haplotypes with WC1 also share those haplotypes with each other or if they share different parts. E.g. Roughly speaking, is the WC1 contribution to Armenians mostly similar to the WC1 contribution to Tajiks, or clearly distinguishable, etc..
@Davidski:
If an individual is getting 20-30% in the GEDmatch "Near East Neolithic K13 Admixture Proportions" calculator's "Iran Neolithic" is that considered significant compared to Near Eastern groups, including any groups mentioned in the paper? I also imagine the calculator will be updated soon...
Strike my last question as I just saw the spreadsheet which is somehow not in excel. Looks like Anatolian Chalcolithic already has that range, and lots of other groups have much above that range, though it does seem that the highest ranging modern populations are similar to those listed in this paper as having more similarity to WC1. There are also some relatively high "Iran Neolithic" modern populations with respect to that calculator that are not listed as sharing special affinity to WC1 (e.g. Assyrian, Iraqi Jews, and of course Kurds). Possibly those populations happened to be more similar to the Iran Neolithic samples but not so much to WC1, or something else...
ReplyDelete