A Revised Model of Anatomically Modern Human Expansions Out of Africa through a Machine Learning Approximate Bayesian...

Area of Science:

• Human Evolution
• Population Genetics
• Paleoanthropology

Background:

• Africa is widely accepted as the origin of anatomically modern humans (AMH).
• The exact dispersal routes and timing of AMHออกจาก Africa remain debated, with single vs. multiple wave hypotheses.
• Previous genetic studies have not definitively resolved these out-of-Africa dispersal models.

Purpose of the Study:

• To test the single vs. multiple dispersal models for anatomically modern humans (AMH) leaving Africa.
• To evaluate the efficacy of Approximate Bayesian Computation with Random Forests for distinguishing population history models.
• To compare simulated genetic data with real genomic variation from modern and archaic populations.

Main Methods:

• Utilized Approximate Bayesian Computation (ABC) coupled with the Random Forest algorithm.
• Simulated genetic data to assess the model's ability to differentiate between single and multiple AMH dispersal scenarios.
• Compared simulation outputs with genomic variation data from contemporary and ancient human populations.

Main Results:

• The study found that a multiple dispersal model is approximately four times more likely than a single dispersal model.
• Estimated two distinct dispersal events: one around 74,000 years ago and another around 46,000 years ago.
• The method successfully distinguished between the alternative out-of-Africa hypotheses.

Conclusions:

• Evidence strongly supports a model of multiple dispersals of anatomically modern humans (AMH) from Africa.
• The findings provide a refined timeline for key human migration events out of Africa.
• This research contributes to understanding the complex genetic history of human colonization.