Genomic data reveal a complex making of humans

Abstract

In the last few years, two paradigms underlying human evolution have crumbled. Modern humans have not totally replaced previous hominins without any admixture, and the expected signatures of adaptations to new environments are surprisingly lacking at the genomic level. Here we review current evidence about archaic admixture and lack of strong selective sweeps in humans. We underline the need to properly model differential admixture in various populations to correctly reconstruct past demography. We also stress the importance of taking into account the spatial dimension of human evolution, which proceeded by a series of range expansions that could have promoted both the introgression of archaic genes and background selection.

Figure 1. Sketches of different scenarios of human dispersal and admixture with archaic human populations during their range expansion out of Africa.

Red arrows indicate approximate migration routes. Neanderthal range is in blue, Denisovan range(s) in orange, and the location of the Denisova site is indicated as a black star. Question marks in the Denisovan range indicate uncertainty on Denisovan hominin presence. Filled ellipses indicate potential places of admixture in scenarios (A–C). (A) Scenario of Reich et al. , with pulses of admixture between modern humans and Neanderthals (dark blue ellipse) and between modern humans and Denisovans (yellow ellipse). (B) Scenario of Rasmussen et al. with two waves into Asia. Denisovan admixture in Oceanians would have occurred during the first wave, possibly at different places during the migration. (C) Scenario of Skoglund and Jakobsson , with distinct Denisovan admixture events in Oceanians and East Asians. (D) Extension of the spatially explicit scenario of Currat and Excoffier postulating a continuous admixture between modern humans and archaic hominins along migration routes overlapping with archaic hominin ranges. Different shades of orange indicate potentially different archaic hominin populations in Asia.

Figure 2. Biased estimation of divergence time and population sizes in case of admixture.

(A) Model of population divergence and admixture: one of two populations having diverged T_Div generations ago has received a fraction a of its genes from another unsampled population that diverged 14,000 generations ago (350,000 y assuming a generation time of 25 y). All populations sizes are assumed to consist of N = 20,000 haploids. (B) Estimated divergence time as a function of initial admixture rate a. (C) Estimated admixed population size for different divergence times and admixture rates. Simulated data consisted of 400,000 segments of 50 bp, thus totalling a 20-Mb DNA sequence. Parameters are estimated by maximizing the probability of the observed joint site frequency spectrum (SFS) , where the expected SFS is estimated by simulation following the approach of Nielsen et al. .

Figure 3. Effect of background selection (BGS) on molecular diversity within and between populations.

After a BGS episode, deleterious mutations (shown in red) are eliminated together with neutral mutations on the same chromosome, leading to reduced diversity. For illustrative purposes, initial neutral diversity is identical in all cases (A–C). Comparison of cases (A) and (B) shows that different BGS episodes will contribute to populations' genetic differentiation. Comparison of cases (B) and (C) shows that recombination reduces the effect of BGS, maintaining diversity, and reducing linkage disequilibrium (LD) as well as population differentiation (compare final states in [A] and [C]).