Hypothesis: A modern human range expansion ∼300,000 years ago had a critical role in Neandertal origins

Abstract

This paper proposes that a critical event in Neandertal evolution was a range expansion of early modern humans into Europe 300 thousand years ago (ka). As bands of early modern humans spread, they incorporated resident archaic people leading ultimately to so much admixture that the resulting population carried ∼3-7% modern ancestry. I identify multiple ways in which this would lead to a high probability of these direct ancestors of Neandertals retaining early modern human lineages at both mitochondrial DNA and Y chromosomes while clustering with Denisovans genome-wide. A previous explanation for these patterns without the unique population dynamics associated with a range expansion has less than a 3% chance of matching the data. In Africa, the hypothesis also has explanatory power: if the expanding population also mixed African archaic humans, it could account for the evidence that all humans living today are admixed between lineages that separated more than a million years ago and remixed ∼300 ka. This hypothesis of a bicontinental range expansion correlates with shared archaeological traits between West Eurasia and Africa that were far less present in East Asia, including Levallois point stone tools, improved hafting, use of fire and ochre, and larger and more organized campsites. Motivated by the evidence that an early modern human expansion had a role in Neandertal and modern human evolution, I identify 66 loci spanning <0.1% of the genome where people today share ancestry with Neandertals more recently than separation from Denisovans. This includes a haplotype at FADS1/FADS2 that codes for an enzyme modifying polyunsaturated plant-derived fatty acids and may reflect adaptation to a more diversified diet, and an amino acid substitution at ARHGAP32 modulating neural connections that may be important in modern human cognitive evolution.

Footnotes

• This new submission reflects extensive new analysis, detailed in four supplementary notes and three supplementary tables. It is changed relative to the original manuscript in three main ways: (1) It is reorganized to be *genetics first*. The original manuscript was presented as a synthesis of genetic and archaeological observations. Rather than starting from a premise that depended on an interpretation of archaeology, I now use entirely genetic arguments to show that models that have been previously used in simulations of the ∼300-200 ka interbreeding of early modern humans and Neandertal ancestors are highly unlikely to fit. But the proposed new model explains disparate observations. I begin by showing that a non-geographically-explicit model that was simulated by Petr et al. Science 2020 is unlikely to explain the observed joint patterns for mtDNA, the Y chromosome, and the autosomes. Given what we know about mtDNA biology, Y chromosome conservation patterns across mammals, and selective constraints, I show that this model has less than a 2-3% probability of being consistent with the data. I then go on to show that geographically explicit simulations of a range expansion accommodate multiple scenarios that predict the data pattern with high probability. Only at the end do I link to the archaeological evidence. (2) Through my new simulation studies, I identify multiple plausible scenarios that would explain the Y chromosome and mtDNA findings with high probability. One is a sex-biased modern human range expansion combined with individuals from matrilines or patrilines tracing their lineages back to modern humans. A second is a modern human range expansion where females stay within the group and where hybrid males have a skew toward at least 60% daughters. A third example is two successive range expansion: one where females stay within the group (spreading modern human lineage mtDNA) and one where males stay within the group (spreading modern human lineage Y). (3) I present a catalog of locations where all living humans and Neandertals share ancestors <625,000 years ago with Denisovans far more distantly related; these previously overlooked loci are candidates for key early modern human innovations. One example is a haplotype at the FADS1 and FADS2 genes which is involved in more efficient conversion of plant-derived fatty acids raising the possibility that a diversified diet could have been adaptive during the early modern human population expansion. Another is at the gene ARHGAP32 which modulates the structure of neural connections.

Funder Information Declared

Howard Hughes Medical Institute, https://ror.org/006w34k90, N/A

National Institutes of Health, HG012287

The copyright holder for this preprint is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license.