Inference of natural selection from ancient DNA

Abstract

Evolutionary processes, including selection, can be indirectly inferred based on patterns of genomic variation among contemporary populations or species. However, this often requires unrealistic assumptions of ancestral demography and selective regimes. Sequencing ancient DNA from temporally spaced samples can inform about past selection processes, as time series data allow direct quantification of population parameters collected before, during, and after genetic changes driven by selection. In this Comment and Opinion, we advocate for the inclusion of temporal sampling and the generation of paleogenomic datasets in evolutionary biology, and highlight some of the recent advances that have yet to be broadly applied by evolutionary biologists. In doing so, we consider the expected signatures of balancing, purifying, and positive selection in time series data, and detail how this can advance our understanding of the chronology and tempo of genomic change driven by selection. However, we also recognize the limitations of such data, which can suffer from postmortem damage, fragmentation, low coverage, and typically low sample size. We therefore highlight the many assumptions and considerations associated with analyzing paleogenomic data and the assumptions associated with analytical methods.

Keywords: Adaptation; ancient DNA; natural selection; paleogenomics; time series.

Figure 1

Complex demographic scenarios in which selective sweeps, due to a novel selection pressure acting upon at least one population from time T _s onward, can be masked or misinterpreted. In each scenario, sampling before (1), during (2), and after (3) T _s provides a time series of allele frequencies in populations A, B, and C, providing more power to infer the true evolutionary history. Allele frequencies are indicated by coloring of branches. i. Positive selection for a derived (red) allele in population B at time T _s drives it to high frequency, differentiating population B from populations A and C, but this differentiation at this locus is later masked by introgression from population B into population C. ii. The same evolutionary history as in i., except this time recent introgression of the ancestral allele (blue) from population A into population B masks the ancestral selection on the derived allele. iii. Parallel selection acts upon the derived allele at time T _s in both populations B and C. Three population selection tests such as the Population Branch Statistic can misinterpret this pattern of differentiation of A from both B and C as that of selection on the ancestral (blue) allele in population A (see Mathieson 2019 for an example of this type of scenario and selection on loci within the FADS gene in humans).

Figure 2

Illustration of how to track genetic adaptation of a population to environmental change through time. (A) One way to catch genetic adaptation in the act is by sampling genetic data of a population before and after the introduction of a new selective pressure (time 2 and 3, respectively). (B) Conceptual illustration of how the frequency of an allele can change in response to a new selective pressure. (C) Significant changes in allele frequencies between different populations (i.e., at time 2 and 3) can be measured with a genome‐wide scan for selection (the figure was created using the gwasResults dataframe included in qqman package in R; Turner 2014).

Figure 3

Theoretical allele trajectories under directional selection for a dominant, additive, and recessive advantageous allele. The fitness (W) of the different genotypes (W ₁₁, W ₁₂, W ₂₂) is defined as W ₁₁ = W ₁₂ > W ₂₂ for a dominant, W ₁₁ > W ₁₂ > W ₂₂ for an additive, and W ₁₁ > W ₁₂ = W ₂₂ for a recessive advantageous allele (allele trajectories were simulated using custom R code; R Core Team 2019).

Figure 4

Illustrative scheme of differences in allele frequencies of an allele under balancing selection versus a neutral allele. An allele under balancing selection (shown in blue) will show small fluctuations around a 0.5 allele frequency. The frequencies of neutral alleles (shown in orange) will change following a more stochastic process, eventually leading to fixation or loss of the allele from the population.