A working model on the peopling of Eurasia

This post will deal with the peopling of Eurasia by modern humans and their subsequent Paleolithic movements and divergence patterns. – Part I of our "in development" series on the question to 'who we are and where we came from': Paleolithic to Epipaleolithic period.

Overview

Modern humans have a "recent African" origin, having evolved within the African continent, and diversified there; one of those branches became isolated in Northeast Africa and would subsequently became ancestral to all modern living 'non-African populations': Eurasians. This branch would carry out the "Out of Africa" migration, – but they were not the first humans to leave Africa; they were preceded by earlier extinct waves of modern humans, possibly distantly related to those of the later main OoA wave – as well as deeper archaic groups which would have become ancestral to the Neanderthals or Denisovans.

During or shortly after the Out of Africa migration of modern humans (c. 60kya), Basal Eurasians (or 'Ancient North Africans') diverged and stayed in Northeast Africa and possibly the Arab peninsula. – The majority of the OoA wave settled in a region encompassing the northern Mesopotamia region and the Persian plateau (c. 55kya), which acted as "population Hub" for the OoA movements (Vallini et al. 2024).

It was in this region, where archaic introgression (Neanderthal admixture) took place. – Subsequently, the Hub population started to develop internal differentiation into a broadly Western Eurasian clade (UP wave) and a broadly Eastern Eurasian clade (IUP wave, including extinct IUP European lineages):

The first and deepest split can be associated with the Zlaty Kun and Ranis lineage in Europe, which is basal to all other Eurasians, and can be described as Crown-like. This lineage headed into Europe, but became extinct without contributing ancestry to later populations. These samples displayed additional Neanderthal admixture.

Around 50kya, the major split between 'Ancient West Eurasians' and 'Ancient East Eurasians' was completed, although there still existed some mutual contact events within the Hub on the Persian plateau; – at least until around 48kya, when Ancient East Eurasians started to expand, in tandem with the affilated 'Initial Upper Paleolithic' material culture. -> IUP wave.

IUP period:

This IUP wave (c. 48kya) is ancestral to all ancient and modern East Eurasian populations. – A northern IUP route gave rise to the IUP microblade-based tool cultures, represented by the BachoKiro_IUP and Oase_IUP specimens in Europe, as well as multiple remains found in Central Asia, Siberia and Northwest China, such as Kara Bom etc., who are all forming a substructured basal sister clade to modern East Eurasians; those IUP lineages largely went extinct, but contributed to the formation of succeeding UP Europeans, as well as to the Tianyuan lineage in Northern China; also evident through the unusual genetic affinity between the European GoyetQ116-1 and the Chinese Tianyuan specimens.

A southern IUP route wave with a coherent genetic makeup ('East Eurasian Core'; "EEC") headed into Southern Asia, where they accumulated shared drift (EEC-drift). After some time, they rapidly diverged into the AASI lineage (ancient South Asian), the Önge lineage (Southeast Asian), the ESEA lineage (broadly Eastern Asian), and the Australasian lineage (Oceanian); ancestral to all modern East Eurasian populations of the Asia-Pacific region; Australasians may harbor an additional pre-EEC component (Oceania_IUP):

The northern and southern route dispersals for the deep East Eurasian branches (IUP_North and EEC_South) are also evident in terms of IUP-affilated material culture:

The IUP and Core & Flake based cultures encountered each other in Northern China, fitting the observed geneflow between IUP and ESEA/Tianyuan lineages:

Both archaeogenetic and archaeologic data support a single southern route dispersal for the EEC, and subsequently rapid divergence in the South–Southeast Asia region:

A single major migration of modern humans into the continents of Asia and Sahul was strongly supported by earlier studies using mitochondrial DNA, the non-recombining portion of Y chromosomes, and autosomal SNP data [42–45]. Ancestral Ancient South Indians with no West Eurasian relatedness, East Asians, Onge (Andamanese hunter–gatherers) and Papuans all derive in a short evolutionary time from the eastward dispersal of an out-of-Africa population [46,47]. The HUGO (Human Genome Organization) Pan-Asian SNP consortium [44] investigated haplotype diversity within present-day Asian populations and found a strong correlation with latitude, with diversity decreasing from south to north. The correlation continues to hold when only mainland Southeast Asian and East Asian populations are considered, and is perhaps attributable to a serial founder effect [50]. These observations are consistent with the view that soon after the single eastward migration of modern humans, East Asians diverged in southern East Asia and dispersed northward across the continent.

Our qpGraph model for Ancient East Eurasians:

Note: it is also possible to model Australasians as just a single distinct EEC branch with extra Denisovan/archaic inputs, next to Önge, AASI and ESEA; e.g. rather than an admixture between basal Önge and an Oceanian IUP. – We decided to use this scenario in regards to the earlier archaeologic findings in Australasia, which contrast the later genetic divergence date estimation (c. 40kya for EEC). – Yet those findings may be affilated with archaic humans, thus a rapidly diversified EEC source could be an simpler explanation.

UP period:

During the IUP expansion, ancestral West Eurasians stayed in the Hub region, accumulating shared genetic drift. At around 41kya, ancestral West Eurasians started to expanded out of the Hub, being affilated with Upper Paleolithic material culture (Aurignacian and Gravettian like tools). These West Eurasian lineages which expansed out of the Hub became known as the 'West Eurasian Core' ("WEC"):

Those heading into Europe would encounter remnants of the East Eurasian-affilated IUP populations, which whom they would partially absorb, and subsequently gave rise to the Aurignacian culture(s) represented by GoyetQ116‐1, and the Gravettian culture(s) represented by Sunghir/Kostenki and the Vestonice cluster.

In Siberia, the UP wave would merge with local EEC groups (Onge/Tianyuan/ESEA-like) to give rise to the Ancient North Siberians and derived Ancient North Eurasians (ANS/ANE). The ANS/ANE have previously been described as "Paleolithic admixture event between WEC and EEC lineages". The ANE/ANS lineage expanded massively throughout Eurasia, mostly via an East to West expansion, contributing to multiple groups such as the Mesolithic "Eastern hunter-gatherers" (EHG) in Europe, the "West Siberian hunter-gatherers" (WSHG), as well as towards populations of the Persian plateau and Southwest Asia, and finally also to the formation of Native Americans (in tandem with a derived ESEA component).

The WEC lineages which populated Anatolia and the Caucasus would give rise to several of the most important West Eurasian lineages: namely the Proto-WHG (which would later merge with UP Europeans as well as some ANE-like geneflow to form the WHG), the Anatolian_HGs (which would give rise to later Anatolian and European Farmers and also contributed significantly to other nearby groups), and the Caucasus_UP as well as Levant_UP groups. Those related to the Levant_UP groups would later (c. 30kya) also expand into Northern Africa, resulting in the formation of the Iberomaurusian culture cline, by merging with the local "basal" Eurasian, specifically 'Ancient North African' lineage (as well as possible deeper African SAHG-like lineages [Aterian remnants?]).

Subsequent geneflow from "basal Eurasian like" groups from Northeast Africa or the Arab peninsula would left some genetic influences onto the Levantine groups (and beyond), resulting in the formation of the Epipaleolothic Natufians, who would again expand throughout the Arab peninsula and Northeast Africa.

The ancestral West Eurasians who stayed in the Hub on the Persian plateau (WEC2), would later, after contact events with "basal" and East Eurasian components, in part via contat to AASI and or via extensive ANE-like geneflow, re-emerge as the Iranian Mesolithic hunter-gatherers and Neolithic Iranians.

In regards to the nature of "Basal Eurasians" and their impact onto West Eurasians:

Based on the archaeogenetic and demographic data presented by Vallini et al. 2024, as well as data on Y chomosome diversity and material culture evidence, we argue that the inferred split between "Basal Eurasians" and "Common Eurasians" happened during the OoA exit, with Basal Eurasians staying in Northeast Africa and on the Arab peninsula.

We first modeled the "basal" component as deepest split on the Proto-Eurasian branch, but after the inferred split of the Gumuz, an East African population without Eurasian geneflow. – Our inferred "Basal" simulation is thus close, if not identical, to the inferred 'Ancient North African' (ANA) lineage (Lazaridis 2018) – There may have been a cline of multiple inter-connected basal lineages. – The one relevant for admixture into later West Asian groups may correspond to the Emiran culture in Northeast Africa and the southern Levant.

A second model differentiated between ANA and Basal Eurasians, with the former staying in Northeast Africa and the later participating in the OoA exit, but diverging soon afterwards, staying on the Arab peninsula.

Basal Eurasian describes a "deep" ancestry without archaic admixture; ANA (Ancient North African) describes a "deep" ancestry "deeper than Basal Eurasians", but a phylogenetic sister lineage to the OoA exit group to the exclusion of other African lineages.

E.g. the supposed "homeland" of the inferred "Basal Eurasians" would be found primarily in Northeast Africa and partly covering the western Arab peninsula; while the Common Eurasian Hub would be found in a region stretching from northern Mesopotamia to the Persian plateau:

To determine the amount of "basal" ancestry, one must be carefull; previous papers often used UP Europeans or the drifted WHG as West Eurasian proxy to determine the amount of needed "basal" ancestry among West Asian or Iranian remains. We caution that this is may be misleading and distorting the real amounts of ancestry components for ancient/modern West Asians.

The reason for that is: Neither UP Europeans nor WHGs can perfectly represent the WEC type for phylogenetically more basal Anatolian/Caucasus_UP or Iranian and Levant groups. While UP Europeans are derived from a shared WEC source in Western Asia/the Persian plateau, those diverged and underwent unique drift to their own compared against the WEC source for Anatolian/Caucasus_UP groups who stayed in West Asia. The divergence between those is most likely >38kya (based on the age of the oldest WEC-affilated sample in Europe: Kostenki14).

Furthermore those UP Europeans (GoyetQ116-1, Sunghir, Muierii, etc.) have variable amounts of preceeding IUP-affilated admixture (BachoKiro_IUP; distantly related to modern East Eurasians). – In turn, the WHG formed from a merger of UP Europeans with significant geneflow from a sister lineage of Anatolian HGs, as well as a small ANE-related component, also evident in the chrY R1b clade among the Villabruna cluster.

In our view, neither UP Europeans nor WHG should be used as proxy for the West Eurasian/WEC ancestry of Caucasus_UP, Iran_N, Natufians, etc., to determine the amount of "Basal" ancestry, as those have non-WEC admixture (BachoKiro_IUP) and or are additionally strongly drifted (WHGs). – Not accounting this, will lead to the need of deeper ancestry among those lacking that admixture; i.e. causing the Basal Eurasian artefact. – E.g. using WHG as WEC proxy is similar in trying to model Onge with Amur_N. -> Onge or Hoabinhian will need more basal ancestry than the derived Amur_N can provide, distorting the results. Using UP Europeans as "unadmixed" proxy is similar to the use of Karitiana/Amerindians as "unadmixed" group in older (2014 studies); etc. It will cause distortions.

-> Previous studies inferred very high amounts of such Basal ancestry, in a range from 45–62%.

We argue that this percentage is significantly too high, caused by the use of wrong imperfect proxies. Simulations and proxies must be chosen by testing those for plausibility.

An example for a possible problematic proxy use is this graph:

E.g. very high "basal" for Natufian and Iran_N while their West Eurasian sources being respectively WHG and EHG; Mesolithic Europeans, which is questionable for West Asian populations, except one follows the scenario in which Mesolithic WHG-like ancestry from Europe is the source for West Eurasian ancestry among West Asians, etc.

As similar case; the previous modeling of Iran_N: two-way models, and or the missing of the newly defined 'WEC2' proxy (Vallini et al. 2024), will results in the need of additional deep admixture to counter the absence of that component.

The Lazaridis 2018 model (ANA+Basal, BUT with UP Europeans as unadmixed, which would fail as those have variable amounts of IUP admixture; e.g. GoyetQ116-1 for example):

A further problem with the above graph is that Mota_HG is unadmixed. We know that Mota_HG has deep admixture (SAHG-like), rather than to have contributed to South African HGs.

Compare to this graph on Iran_N, excluding the supposed 'WEC2' proxy, but allowing Anatolia/Caucasus_UP to be unadmixed WEC:

In this cases, Iran_N (Ganj_Dareh_N) needs significantly increased EEC-like (or Tianyuan-like) ancestry (D1; D3) to counter the absence of the 'WEC2' component (first defined by Vallini et al. 2024), which is a plausible component for the Paleolithic/Mesolithic Iranian populations:

We simulated two different West Eurasian populations: WEC and WEC2, with WEC2 staying in the Hub longer than WEC (and Kostenki14), and hence closer to it from a genetic point of view. [...] Our results showed that the genetic component closest to the Hub population is represented in ancient and modern populations in the Persian Plateau. Such a component, after mixing with Basal and East Eurasian ancestries, resurfaced in the palaeogenetic record, previously referred to as the Iranian Neolithic, the Iranian Hunter Gatherer’ or the East Meta49.

Lazaridis 2018 (Supplementary Information) already noted the possibility that Iran_N may have an ancestry component similar to the WEC2 component defined by Vallini et al. 2024.

We argue that by applying the scenario presented by Vallini et al. 2024, we get a more accurate model on these broadly West Eurasian populations.

Our qpGraph models for Ancient West Eurasian lineages:

ANA only:

ANA+Basal:

We argue that Anatolia_UP and Caucasus_UP lineages do not need significant amounts of "deep/Basal" ancestry, rather they can be modeled as "nearly unadmixed" representatives of the local WEC lineage (closer to "basal WEC"), while derived UP Europeans display variable amounts of pre-existing IUP-affilated admixture. This is inline with the repetive expansions from the population Hub, as well as the pre-existing IUP cultures within Europe. Similarly, WHG display some IUP and ANE components via their demographic formation during the early Mesolithic (Epigravettian) period.

Henceforth, "basal" admixture is decreased albeit not absent from ancient and modern West Asians, specifically the Anatolian/Caucasus_UP lineage.

The remainder "basal" ancestry needed for Iran_N and Natufians can best be represented by geneflow from either Northern Africa only (e.g. ANA only = technically Basal Eurasian; a sister lineage of Eurasians which stayed in Africa) or a combination of ANA and Basal (North Africa and Arabia). – E.g. geneflow from Northern Africa and or Arabia.

A yet imperfect proxy for ANA may be the non-WEC component among Mota_HG or among Iberomaurusians (but excluding the small respective deep SAHG-like admixture/affinity [especially for Mota_HG]). – For Basal Eurasians (if they existed), we currently have no archaeologic nor archaeogenetic remains which could be affilated with such 'Basal Arabia' lienage.

Our results fit with the models presented by Vallini et al. 2024. For a detailed estimation of Basal (BEA), WEC/WEC2 and IUP/EEC amounts for ancient and modern populations, see supplementary Data 1–15; 11.

We argue that this refined model viz a viz the repetive population Hub (IUP ; UP waves) can explain the complex demographic events in Eurasia during the Paleolithic period better than more simplistic models, or models in which UP Europeans are "unadmixed", which is at odds with the presence of multiple non-West Eurasian groups, especially the pre-existing IUP groups, which got absorbed by expanding WEC lineages from Western Asia/the Persian plateau.

Summary

The peopling of Eurasia was plausibly carried out via a repetive population Hub (on the Persian plateau).

The first successful major expansion was the IUP/EEC wave (Ancient East Eurasians), which populated most of Eurasia; with the EEC heading into the South–Southeast Asia region, diversifying there, and the main source of ancestry for all modern East Eurasians.

The second successful major expansion was the UP/WEC wave (Ancient West Eurasians), which expanded after the IUP wave. The UP wave partly absorbed and partly replaced the previous IUP groups. The deep West Eurasian component which stayed in the Hub region (WEC2) contributed primarily to later Mesolithic and Neolithic Iranian groups.

The phylogenetic patters show that the IUP expansion gave rise to the northern route IUP lineage(s) and the southern route EEC lineage(s); the Ust'Ishim lineage similarly expanded during the IUP wave, but shares only minimal evolutionary drift with the rest of the IUP/EEC lineage, pointing to a slightly early divergence than the main wave ancestral to BachoKiro_IUP,  Oase_IUP or the EEC (see also Vallini et al. 2022 or Allentoft et al. 2024 supplementary informationen).

"The Ust’Ishim lineage is described as "near trifurcation" between West and East Eurasians, but sharing a short period of evolutionary drift with Eastern Eurasians, having diverged from their ancestor shortly after the divergence from Ancient Western Eurasians."

The UP wave gave rise to the WEC lineage(s) expanding from the Hub, and the WEC2 lineage which stayed within the Persian plateau.

These complex patterns of expansions, migration and contact events shaped the Paleolithic to Epipaleolithic Eurasian demographic makeup. Outgoing from these newly emerged genetic lineages and further migration and contact events, but also isolation events, modern Eurasian sub-populations formed.

For comparison, here the often cited scenario with high "Basal Eurasian" ancestry and UP Europeans as unadmixed; by Allentoft et al. 2024:

In this scenario, UP Europeans represent the source of West Eurasian ancestry for all other groups, with Caucasus_UP/Anatolians being derived and admixed with more basal ancestry, while Iran_N is a merger of local Basal Eurasian Gulf populations and Ancient North Eurasian (Paleolithic mix of WEC and EEC) geneflow:

While we can not exclude the possibility for the above scenario, we argue that the Vallini scenario of the repetive population Hub is more likely, and better aligns with the available archaeogenetic and archaeologic data.

Finally, a compilation of our graphical representation of the IUP wave and the later UP wave according to the repetive population Hub model, as well as the respective demographic patterns:

We hope our summary on the Paleolithic patterns of the peopling of Eurasia help to get a more compact overview, and provide future directions for entangling the genetic and demographic history of Eurasia and beyond.