Villabruna cluster in Late Epigravettian Sicily supports South Italian corridor for R1b-V88


New preprint Late Upper Palaeolithic hunter-gatherers in the Central Mediterranean: new archaeological and genetic data from the Late Epigravettian burial Oriente C (Favignana, Sicily), by Catalano et al. bioRxiv (2019).

Interesting excerpts (emphasis mine):

Grotta d’Oriente is a small coastal cave located on the island of Favignana, the largest (~20 km2) of a group of small islands forming the Egadi Archipelago, ~5 km from the NW coast of Sicily.

The Oriente C funeral pit opens in the lower portion of layer 7, specifically sublayer 7D. Two radiocarbon dates on charcoal from the sublayers 7D (12149±65 uncal. BP) and 7E, 12132±80 uncal. BP are consistent with the associated Late Epigravettian lithic assemblages (Lo Vetro and Martini, 2012; Martini et al., 2012b) and refer the burial to a period between about 14200-13800 cal. BP, when Favignana was connected to the main island (Agnesi et al., 1993; Antonioli et al., 2002; Mannino et al. 2014).

A-B) Geographic location of Grotta d’Oriente.

The anatomical features of Oriente C are close to those of Late Upper Palaeolithic populations of the Mediterranean and show strong affinity with other Palaeolithic individuals of Sicily. As suggested by Henke (1989) and Fabbri (1995) the hunter-gatherer populations were morphologically rather uniform.

Genetic analysis

We confirmed the originally reported mitochondrial haplogroup assignment of U2’3’4’7’8’9. This haplogroup is present in both pre- and post-LGM populations, but is rare by the Mesolithic, when U5 dominates (Posth et al.2016).

Lipson et al. (2018) (their supplementary Figure S5.1) and Villalba-Mouco et al. (2019) (their Figure 2A) showed that European Late Palaeolithic and Mesolithic hunter-gatherers fall along two main axes of genetic variation. Multidimensional scaling (MDS) of f3-statistics shows that these axes form a “V” shape (Fig. 3). (…)

Focusing further on Oriente C, we find that it shares most drift with individuals from Northern Italy, Switzerland and Luxembourg, and less with individuals from Iberia, Scandinavia, and East and Southeast Europe (Fig. 4A-B). Shared drift decreases significantly with distance (Fig. 4C) and with time (Fig. 4D) although in a linear model of drift with distance and time as a covariate, only distance (p=1.3×10-6) and not time (p=0.11) is significant. Consistent with the overall E-W cline in hunter-gatherer ancestry, genetic distance to Oriente C increases more rapidly with longitude than latitude, although this may also be affected by geographic features. For example, Oriente C shares significantly more drift with the 8,000 year-old 1,400 km distant individual from Loschbour in Luxembourg (Lazaridis et al.,2014), than with the 9,000 year old individual from Vela Spila in Croatia (Mathieson et al.,2018) only 700 km away as shown by the D-statistic (Patterson et al.,2012) D (Mbuti, Oriente C, Vela Spila, Villabruna); Z=3.42. Oriente C’s heterozygosity was slightly lower than Villabruna (14% lower at 1240k transversion sites), but this difference is not significant (bootstrap P=0.12).

Multidimensional scaling of outgroup f3-statistics for Late 531 Upper Palaeolithic and Mesolithic hunter-gatherers.

Discussion and Conclusion

The robust record of radiocarbon dates proves that they reached Sicily not before 15-14 ka cal. BP, several millennia after the LGM peak. In our opinion, in fact, the hypothesis about an early colonization of Sicily by Aurignacians (Laplace, 1964; Chilardi et al., 1996) must be rejected, on the basis of a recent reinterpretation of the techno-typological features of the lithic industries from Riparo di Fontana Nuova (Martini et al., 2007; Lo Vetro and Martini, 2012; on this topic see also Di Maida et al., 2019).

These analyses have implications for understanding the origin and diffusion of the hunter-gatherers that inhabited Europe during the Late Upper Palaeolithic and Mesolithic. Our findings indicate that Oriente C shows a strong genetic relationship with Western European Late Upper Palaeolithic and Mesolithic hunter-gatherers, suggesting that the “Western hunter-gatherers” was a homogeneous population widely distributed in the Central Mediterranean, presumably as a consequence of continuous gene flow among different groups, or a range expansion following the LGM.

The same statistic as in A plotted with geographic position

The South Italian corridor

Once again, a hypothesis based on phylogeography – apart from scarce archaeological and palaeolinguistic data (“Semitic”-like topo-hydronymy and substrates in Europe) – seems to be confirmed step by step. Since the finding of the Villabruna individual of hg. R1b-L754 (likely R1b-V88, like south-eastern European lineages expanded with WHG ancestry), it was quite likely to find out that southern Europe was the origin of the expansion of R1b-V88 into Africa.

The most likely explanation for the presence of “archaic” R1b-V88 subclades among modern Sardinians was, therefore, that they represented a remnant from a Late Upper Palaeolithic/Early Mesolithic population that had not been replaced in subsequent migrations, and thus that the migration of these lineages into Northern Africa and the Green Sahara happened during a period when Italy was connected by a shallower Mediterranean (and more land connections) to Northern Africa.

Likely Late Epigravettian/Mesolithic expansion of R1b-V88 into Northern Africa. See full map.

Nevertheless, the arguments for a quite recent expansion of R1b-V88 through the Mediterranean and into Africa keep being repeated, probably based on ancestry from the few ancient (and many modern) populations that have been investigated to date, a simplistic approach prone to important errors that overarch whole migration models.

For example, in the recent paper by Marcus et al. (2019) the presence of these lineages among ancient Sardinians (from the late 4th millennium BC on) is interpreted as an expansion of R1b-V88 with the Cardial Neolithic based on their ancestry, disregarding the millennia-long gap between these samples and the presence of this haplogroup in Palaeolithic/Mesolithic Northern Iberia and Northern Italy, and the comparatively much earlier splits in the phylogenetic tree and dispersal among African populations.

Afroasiatic and Nostratic

I was asked recently if I really believed that we could reconstruct Proto-Nostratic and connect it with any ancestral population. My answer is simple: until the Chalcolithic – when the whole picture of Indo-Europeans, Uralians, Egyptians or Semites becomes quite clear – we have just very few (linguistic, archaeological, genetic) dots which we would like to connect, and we do so the best we can. The earlier the population and proto-language, the more difficult this task becomes.

NOTE. 1) I tentatively connected hg. R with Nostratic in a previous text – when it appeared that R1a expanded from around Lake Baikal, hence Eurasiatic; R1b from the south with AME-WHG ancestry, hence Afroasiatic; and R2 with Dravidian.

2) After that, I though it was more likely to be connected to AME ancestry and the Middle East, because of the apparent expansion of WHG from south-eastern Europe, and the potential association of Afroasiatic and (Elamo-?)Dravidian to Middle Eastern populations.

3) However, after finding more and more R1b samples expanding through northern Eurasia, spreading through the (then wider) steppe regions; and R1a essentially surviving among other groups in eastern Europe for thousands of years without being associated to significant migrations (like, say, hg. C after the Palaeolithic), it didn’t seem like this division was accurate, hence my most recent version.

But, in essence, it’s all about connecting the dots, and we have very few of them…

Phylogenetic tree from Pagel et al. (2013), partially in agreement with Kortlandt’s view on Eurasiatic. “Consensus phylogenetic tree of Eurasiatic superfamily (A) superimposed on Eurasia and (B) rooted tree with estimated dates of origin of families and of superfamily. (A) Unrooted consensus tree with branch lengths (solid lines) shown to scale and illustrating the correspondence between the tree and the contemporary north-south and east-west geographical positions of these language families. Abbreviations: P (proto) followed by initials of language family: PD, proto-Dravidian; PK, proto-Kartvelian; PU, proto-Uralic; PIE, proto–Indo-European; PA, proto-Altaic; PCK, proto–Chukchi-Kamchatkan; PIY, proto–Inuit-Yupik. The dotted line to PIY extends the inferred branch length into the area in which Inuit-Yupik languages are currently spoken: it is not a measure of divergence. The cross-hatched line to PK indicates that branch has been shortened (compare with B). The branch to proto-Dravidian ends in an area that Dravidian populations are thought to have occupied before the arrival of Indo-Europeans (see main text). (B) Consensus tree rooted using proto-Dravidian as the outgroup. The age at the root is 14.45 ± 1.75 kya (95% CI = 11.72–18.38 kya) or a slightly older 15.61 ± 2.29 kya (95% CI = 11.72–20.40 kya) if the tree is rooted with proto-Kartvelian. The age assumes midpoint rooting along the branch leading to proto-Dravidian (rooting closer to PD would produce an older root, and vice versa), and takes into account uncertainty around proto–Indo-European date of 8,700 ± 544 (SD) y following ref. 35 and the PCK date of 692 ± 67 (SD) y ago.”

In linguistics, I trust traditional linguists who tend to trust other more experimental linguists (like Hyllested or Kortlandt) who consider that – in their experience – an Indo-Uralic and a Eurasiatic phylum can be reconstructed. Similarly, linguists like Kortlandt are apparently (partially) supportive of attempts like that of Allan Bomhard with Nostratic – although almost everyone is critic of the Muscovite school‘s attachment to the Brugmannian reconstruction, stuck in pre-laryngeal Proto-Indo-Anatolian and similar archaisms.

I mostly use Nostratic as a way to give a simplistic ethnolinguistic label to the genetically related prehistoric peoples whose languages we will probably never know. I think it’s becoming clear that the strongest connection right now with the expansion of potential Eurasiatic dialects is offered by ANE-related populations (hence Y-chromosome bottlenecks under hg. R, Q, probably also N), however complicated the reconstruction of that hypothetic community (and its dialectalization) may be.

Therefore, the multiple expansions of lineages more or less closely associated to ANE-related peoples – like R1b-V88 in the case of Afrasian, or R2 in the case of Dravidians – are the easiest to link to the traditionally described Nostratic dialects and their highly hypothetic relationship.

Reconstruction of North African vegetation during past green Sahara periods. Estimated and reconstructed MAP for the Holocene GSP (6–10 kyr BP) projected onto a cross-section along the eastern Sahara (left panel) and map view of reconstructed MAP, vegetation and physiographic elements [7,8,11,45] (right panel). Image from Larrasoaña et al. (2013).

What should be clear to anyone is that the attempt of many modern Afroasiatic speakers to connect their language to their own (or their own community’s main) haplogroups, frequently E and/or J, is flawed for many reasons; it was simplistic in the 2000s, but it is absurd after the advent of ancient DNA investigation and more recent investigation on SNP mutation rates. R1b-V88 should have been on the table of discussions about the expansion of Afroasiatic communities through the Green Sahara long ago, whether one supports a Nostratic phylum or not.

The fact that the role of R1b bottlenecks and expansions in the spread of Afroasiatic is usually not even discussed despite their likely connection with the most recent population expansions through the Green Sahara fitting a reasonable time frame for Proto-Afroasiatic reconstruction, a reasonable geographical homeland, and a compatible dialectal division – unlike many other proposed (E or J) subclades – reveals (once again) a lot about the reasons behind amateur interest in genetics.

Just like seeing the fixation in (and immobility of) recent writings about the role of I1, I2, or (more recently) R1a in the Proto-Indo-European expansion, R1b with Vasconic, or N1c with Proto-Uralic.

NOTE. That evident interest notwithstanding, it is undeniable that we have a much better understanding of the expansions of R1b subclades than other haplogroups, probably due in great part to the easier recovery of ancient DNA from Eurasia (and Europe in particular), for many different – sociopolitical, geographical, technological – reasons. It is quite possible that a more thorough temporal transect of ancient DNA from the Middle East and Africa might radically change our understanding of population movements, especially those related to the Afroasiatic expansion. I am referring in this post to interpretations based on the data we currently have, despite that potential R1b-based bias.


Waves of Palaeolithic ANE ancestry driven by P subclades; new CWC-like Finnish Iron Age

New preprint The population history of northeastern Siberia since the Pleistocene, by Sikora et al. bioRxiv (2018).

Interesting excerpts (emphasis mine; most internal references removed):

ANE ancestry

The earliest, most secure archaeological evidence of human occupation of the region comes from the artefact-rich, high-latitude (~70° N) Yana RHS site dated to ~31.6 kya (…)

The Yana RHS human remains represent the earliest direct evidence of human presence in northeastern Siberia, a population we refer to as “Ancient North Siberians” (ANS). Both Yana RHS individuals were unrelated males, and belong to mitochondrial haplogroup U, predominant among ancient West Eurasian hunter-gatherers, and to Y chromosome haplogroup P1, ancestral to haplogroups Q and R, which are widespread among present-day Eurasians and Native Americans.

Symmetry tests using f4 statistics reject tree-like clade relationships with both Early West Eurasians (EWE; Sunghir) and Early East Asians (EEA; Tianyuan); however, Yana is genetically closer to EWE, despite its geographic location in northeastern Siberia

Using admixture graphs (qpGraph) and outgroup-based estimation of mixture proportions (qpAdm), we find that Yana can be modelled as EWE with ~25% contribution from EEA

Among all ancient individuals, Yana shares the most genetic drift with Mal’ta, and f4 statistics show that Mal’ta shares more alleles with Yana than with EWE (e.g. f4(Mbuti,Mal’ta;Sunghir,Yana) = 0.0019, Z = 3.99). Mal’ta and Yana also exhibit a similar pattern of genetic affinities to both EWE and EEA, consistent with previous studies.The ANE lineage can thus be considered a descendant of the ANS lineage, demonstrating that by 31.6 kya early representatives of this lineage were widespread across northern Eurasia, including far northeastern Siberia.


Ancient Palaeosiberian

(…) the 9.8 kya Kolyma1 individual, representing a group we term “Ancient Paleosiberians” (AP). Our results indicate that AP are derived from a first major genetic shift observed in the region. Principal component analysis (PCA), outgroup f3-statistics and mtDNA and Y chromosome haplogroups (G1b and Q1a1a, respectively) demonstrate a close affinity between AP and present-day Koryaks, Itelmen and Chukchis, as well as with Native Americans.

For both AP and Native Americans, ANS ancestry appears more closely related to Mal’ta than Yana, therefore rejecting a direct contribution of Yana to later AP or Native American groups.

Lake Baikal Neolithic – Bronze Age

(…) the newly reported genomes from Ust’Belaya and recently published neighbouring Neolithic and Bronze Age sites show a succession of three distinct genetic ancestries over a ~6 ky time span. The earliest individuals show predominantly East Asian ancestry, closely related to the ancient individuals from DGC. In the early Bronze Age (BA), we observe a resurgence of AP ancestry (up to ~50% ancestry fraction), as well as influence of West Eurasian Steppe ANE ancestry represented by the early BA individuals from Afanasievo in the Altai region (~10%) This is consistent with previous reports of gene flow from an unknown ANE-related source into Lake Baikal hunter-gatherers.

Our results suggest a southward expansion of AP as a possible source, which is also consistent with the replacement of Y chromosome lineages observed at Lake Baikal, from predominantly haplogroup N in the Neolithic to haplogroup Q in the BA. Finally, the most recent individual from Ust’Belaya, dated to ~600 years ago, falls along the Neosiberian cline, similar to the ~760 year-old ‘Young Yana’ individual from northeastern Siberia, demonstrating the widespread distribution of Neosiberian ancestry in the most recent epoch.

Genetic structure of ancient northeast Siberians. PCA of ancient individuals projected onto a set of modern Eurasian and American individuals. Abbreviations in group labels: UP – Upper Palaeolithic; LP – Late Palaeolithic; M – Mesolithic; EN – Early Neolithic; MN – Middle Neolithic; LN – Late Neolithic; EBA – Early Bronze Age; LBA – Late Bronze Age; IA – Iron Age; PE – Paleoeskimo; MED – Medieval

Finland Saami

At the western edge of northern Eurasia, genetic and strontium isotope data from ancient individuals at the Levänluhta site documents the presence of Saami ancestry in Southern Finland in the Late Holocene 1.5 kya. This ancestry component is currently limited to the northern fringes of the region, mirroring the pattern observed for AP ancestry in northeastern Siberia. However, while the ancient Saami individuals harbour East Asian ancestry, we find that this is better modelled by DGC rather than AP, suggesting that AP influence was likely restricted to the eastern side of the Urals. Comparison of ancient Finns and Saami with their present-day counterparts reveals additional gene flow over the past 1.6 kya, with evidence for West Eurasian admixture into modern Saami. The ancient Finn from Levänluhta shows lower Siberian ancestry than modern Finns .

EDIT (27 OCT 2018): By comparing the three, I see these are samples published already (at least two) in Lamnidis et al. (2018), but here with added (1) specific radiocarbon dates, (2) comparison with Neosiberian populations and (3) strontium isotope analyses.

Finnish_IA (ca. 350 AD) is probably a Saami-speaking individual, just like the Saami_IA with newly reported radiocarbon dates from Levänluhta ca. 400-600 AD (since Fennic peoples were then likely around the Gulf of Finland).

The conflicting strontium isotope data on marine dietary resources on certain samples from the supplementary material hint at possible external origin of the diet of some of the previously reported (and possibly one newly reported) Saami Iron Age individuals, from some 25-30 km. to the northwest through the river up to hundreds of km. to the southwest of Levänluhta (i.e. the whole coast of the Bothnian Sea). It is unclear why they would prefer an origin of the dietary source in southern Baltic regions instead of some km. to the west, though, unless that’s what they want to propose based on the sample’s admixture…

The coast of the Bothnian Sea (=the northern part of the Baltic Sea, between Sweden and Finland) lay only 25-30 km to the northwest, and accessible to the Iron Age people of the Levänluhta region via the Kyrönjoki river. (…) For individual JA2065/DA236, the low 87Sr/86Sr value (0.71078) would imply an exceptionally heavy reliance on Baltic Sea resources. The δ13C and δ15N values of the individual are near comparable (especially considering within-Baltic latitudinal gradients in δ13C; Torniainen et al. 2017) to the δ13C and δ15N values of a Middle Neolithic population on the Baltic island of Gotland (Eriksson, 2004) interpreted to have subsisted primarily on seals.

These new data on the samples give us some more information than what we already had, because the early date of Finnish_IA implies that there was few East Asian admixture (if any at all) in west Finland during the Roman Iron Age, which pushes still farther forward in time the expected appearance of Siberian ancestry among Saamic (first) and Fennic populations (later). It is unclear whether this East Asian ancestry found in Finnish_IA is actually related to DGC, or it is rather related to the ENA-like ancestry found already in Baltic hunter-gatherers (i.e. in some EHG samples from Karelia), for which Baikal_EN is a good proxy in Lazaridis et al. (2018).

Since Bronze Age and Iron Age samples from Estonia show more Baltic_HG drift compared to Corded Ware samples, it is likely that this supposedly DGC-related ancestry (here considered part of the ‘Siberian ancestry’) is actually an EHG-related ENA component of north-east European hunter-gatherers, with whom Finno-Saamic peoples admixed during the expansion of the Corded Ware culture into Finland.

The paper finds thus increased (probably the actual) Siberian ancestry in modern Finns compared to this Iron Age Saami individual. Coupled with the later Saami Iron Age samples, from between one to three centuries later – showing the start of Siberian ancestry influx – , we can begin to establish when the expansion of Siberian ancestry happened in central Finland, and thus quite likely when the Saami began to expand to the north and east and admix with Palaeo-Laplandic peoples.

Admixture modelling using qpAdm. Maps showing locations and ancestry proportions of ancient (left) and modern (right) groups.

One sample of haplogroup N1a1a1a1a4a1-M1982, Yana_MED, is found in the Arctic region (north-eastern Yakutia) ca. 1100 AD. Since it is derived from N1a1a1a1a-L392, it might be a surprise for some to find it in a clearly non-Uralic speaking environment at the same time other subclades of this haplogroup were admixing in the west with well-established Finno-Saamic, Volga-Finnic, Ugric, and Samoyedic populations…

On the growing doubts that these data – contradicting the CWC=IE theory – are creating among geneticists (from the supplementary materials):

NOTE. This paper comes from the Copenhagen group, also signed by Kristiansen, one of today’s strongest supporters of this connection

The Proto-Saami language evolved in southern Finland and Karelia in the Early Iron Age, an area now host to Finnish and the closely related Karelian, but with Saami toponyms showing that the latter two languages are intrusive here (Saarikivi 2004). Saami-speaking populations are thought to have retreated to Lapland during the Middle Iron Age (300–800 AD), where it diverged into the modern Saami dialects. Genetically, the northward retreat of the Saami language correlates with the documented decrease of Saami ancestry in Southern Finland between the Iron Age and the modern period (cf. Lamnidis et al. 2018).

On the way to Lapland, the Saami replaced at least two linguistically obscure groups. This can be inferred from 1) an influx of non-Uralic loanwords into Proto-Saami in the Finnish Lakeland area, and 2) an influx of non-Uralic, non-Germanic words into Saami dialects in Lapland (Aikio 2012). Both of these borrowing events imply contact with non-Saami-speaking groups, e.g. non-Uralic-speaking hunter-gatherers that may have left a genetic and linguistic footprint on modern Saami populations.

The linguistic prehistory of Finland thus does not allow for a straightforward interpretation of the genetic data. The detection of East Asian ancestry in the genetically Saami individual is indicative of a population movement from the east (cf. Lamnidis et al. 2018, Rootsi et al. 2007), one that given the affinities with the ~7.6 ky old individuals from the Devil’s Gate Cave may have been a western extension of the Neosiberian turnover. However, it remains unclear whether this gene flow should be associated with the arrival of Uralic speakers, thus providing further support for a Uralic homeland in Eastern Eurasia, or with an earlier immigration of pre-Uralic, so-called “Paleo-Lakelandic” groups.

I think the genetic interpretation is already straightforward, though. We had a sneak peek at how this late admixture with non-Uralians (mainly Palaeo-Lakelandic and Palaeo-Laplandic peoples from Lovozero and related asbestos ware cultures) is going to unfold among expanding Saami-speaking populations thanks to Lamnidis et al. (2018):

PCA plot of 113 Modern Eurasian populations, with individuals from this study projected on the principal components. Uralic speakers are highlighted in light purple. Image modified from Lamnidis et al. (2018)

Also, still no trace of R1a in far East Asia (reported as M17 ca. 5300 BC near Lake Baikal by Moussa et al. 2016), so I still have doubts about my previous assessment that R1a split into M17 (and thus also M417) in Siberia, with those expanding hunter-gatherer pottery.