genomic origins first farmers


Regular Member
Reaction score
The genomic origins of the world’s first farmers

Author links open overlay panelNinaMarchi1222LauraWinkelbach322IlektraSchulz2422MaximeBrami322ZuzanaHofmanová2456JensBlöcher3Carlos S.Reyna-Blanco24YoanDiekmann37AlexandreThiéry1224AdamandiaKapopoulou12VivianLink24ValériePiuz1SusanneKreutzer325Sylwia M.Figarska3ElissavetGaniatsou8AlbertPukaj3Travis J.Struck9Ryan N.Gutenkunst9LaurentExcoffier122326 rights and content
Under a Creative Commons license
Open access


•European HGs diverged from SW Asian HGs during the LGM
•Low genetic diversity of European HGs is due to a strong LGM demographic bottleneck
•Ancestors of western early farmers emerged after repeated post-LGM admixtures
•EFs strongly diverged from SW Asians during their expansion through Anatolia


The precise genetic origins of the first Neolithic farming populations in Europe and Southwest Asia, as well as the processes and the timing of their differentiation, remain largely unknown. Demogenomic modeling of high-quality ancient genomes reveals that the early farmers of Anatolia and Europe emerged from a multiphase mixing of a Southwest Asian population with a strongly bottlenecked western hunter-gatherer population after the last glacial maximum. Moreover, the ancestors of the first farmers of Europe and Anatolia went through a period of extreme genetic drift during their westward range expansion, contributing highly to their genetic distinctiveness. This modeling elucidates the demographic processes at the root of the Neolithic transition and leads to a spatial interpretation of the population history of Southwest Asia and Europe during the late Pleistocene and early Holocene.

Ancestors of western EFs admixed twice with western HGs

We find that the ancestors of western EFs received a second pulse of gene flow (15%, 95% CI 6–17) from the Western metapopulation ∼12.9 kya (95% CI 9.4–13.9), while Caucasus HGs did not (Figure M1_20B). Models that do not include this additional admixture have a lower likelihood and are therefore rejected (Figure M1_20A). Thus, the ancestors of western EFs are the product of repeated episodes of gene flow from the Western metapopulation. These populations have then diverged from Caucasus HGs due to an intense period of genetic drift between 12.9 and 9.1 kya (Figures 3 and 4). Indeed, we find that their effective population size was reduced to 620 individuals (95% CI 72–2,150) during this relatively long period of drift, which caused them to not only diverge genetically from their ancestral population but also from Caucasus and European HGs, and from Iranian EFs (Figure 4).

A last glacial maximum divergence between Eastern and Western metapopulations

Our model also provides important insights regarding the deep branching of pre-Neolithic populations. The divergence between the ancestors of the Western and Eastern metapopulations is estimated to date back to ∼25.6 kya (95% CI 17.3–31.3, Figure M1_18). This is much younger than the previously inferred divergence time between the ancestors of western European HGs and either Iranian EFs (46–77 kya; Broushaki et al., 2016) or European EFs (46 kya; Jones et al., 2015). However, these previous divergence times were obtained using relatively simple models without bottlenecks and assuming topologies with only recent or even no admixture at all.
We have explored additional scenarios to evaluate the effects of these simplifications on metapopulation divergence times. As expected, a model without bottlenecks on the metapopulation branches leads to a much older divergence time of 39 kya between Eastern and Western metapopulations (Table S4), which is more in line with previous estimates, but this model is inherently less likely than the original one (Figure M1_18A). On the other hand, models without any admixture between the Western HG metapopulation and the ancestors of western EFs lead to a much younger divergence time between the Eastern and Western HG metapopulations (16 kya, Table S4), but the fit with the data is very poor (Figure M1_18A).
Comparing two western European HGs from the sites of Bichon and Loschbour with our newly sequenced Mesolithic individuals from Vlasac further reveals that European HG populations had already split during the last glacial maximum (LGM) ∼22.8 kya (95% CI 16.7–24.7, Figure M1_28), with Bichon and Loschbour populations subsequently diverging from one another, approximately 1,000 years later.
The reduced diversity in European HGs is due to a massive LGM bottleneck

Genetic diversity as quantified by the heterozygosity at neutral sites is much lower in HGs than in EFs (Figure 2C), excepting Northwest Anatolian EFs in line with previous studies (Kılınç et al., 2016; Kousathanas et al., 2017). HG genomes furthermore show a generally larger proportion of intermediate runs of homozygosity (ROHs) (2–10 Mb ROHs, Figure 2D; Figure M1_7; Ringbauer et al., 2021) indicative of background relatedness within European HGs, usually attributed to small population size (Ceballos et al., 2021)—a small population size is also observed in MSMC2 analyses (Figure S2).

A split of European HGs triggered by the LGM

Our model suggests that European HGs had already split into two subgroups (West 1 and West 2 in Figure 5C) ∼23 kya, after experiencing a very severe bottleneck during the LGM, responsible for their low level of genetic diversity (Figure 2C). In contrast to previous studies (Ceballos et al., 2021; Günther et al., 2018), HGs were found to have generally larger effective population sizes than contemporary EFs (Figure 3). Such relatively large effective population sizes can lead to slow population differentiation, which might explain why the different HG groups show close genetic affinities (Figures 2A and 4A) despite long divergence times and a wide geographic distribution. Large HG effective population sizes could be due to long-distance genetic exchanges between groups. Contrastingly, the inferred low effective population size of EFs (despite obvious large census sizes) suggests that the Neolithic transition was linked to a reduction in local EF effective population sizes, potentially due to “sedentarization” or commitment to place (Aimé et al., 2013) and restricted gene flow among small-scale farming communities, as observed at the aceramic Neolithic sites of Boncuklu and Aşıklı (Yaka et al., 2021).


Although the world’s first farmers look genetically very different from European HGs, our simulation based on high-quality genomes shows that some European and Southwest Asian populations in fact shared a recent common history marked by repeated interactions since the end of the last ice age. Strong drift during their expansion through Anatolia contributed to making western EFs look more dissimilar than they actually were and somehow concealed their hybrid nature. In summary, the idea of a single cultural and genetic origin of all farmers in the Fertile Crescent, without significant initial contribution of European-like HGs, is no longer tenable in its current form.
MDS seems to be a new fashion.
However I have the impression that 15 genomes, of which the oldest is abt 13 ka offers enough resolution for a detailed 26.000 year old history.
WHG would have split from CHG only 25,6 ka, but as far as I remember 26 ka CHG had 28 % Dzudzuana ancestry and WHG none.
correction : Dzudzuana = 72 % West Eurasian + 28 % Basal Eurasian
Anyway, the pedigree in this study is not compatible with the admix tree of the Dzudzuana paper.
WHG would have split from CHG only 25,6 ka

I think I get what this means, the older date for European and Middle Eastern Hunter-gatherers divergence was some 45,000 years ago. Which is a similar date for U5 and IJ haplogroups separation.

But this new study seem to detect further admixture among hroups that leave their effective branching, much more recent in time.

This thread has been viewed 2711 times.