Sexual selection, genomic evolution and population fitness in Drosophila pseudoobscura

Abstract

Sexual selection shapes the genome in unique ways. It is also likely to have significant fitness consequences, such as purging deleterious mutations from the genome or conversely maintaining genetic load in a population via sexual conflict. Here, we examined what the influence of sexual selection has on genomic variation potentially underlying population fitness using experimentally evolved Drosophila pseudoobscura populations. Sexual selection was manipulated by keeping replicate lines in elevated polyandry or strict monogamy for approximately 200 generations followed by individual-based sequencing. Using pi (π), fixation index (F_st)and recombination rate measures, we confirmed signatures of selection were not dispersed but mainly localized to the third and X chromosome. Overall mutational load was similar between lines but our analysis of the distribution of fitness effects revealed considerable variation between lines and chromosomes. Furthermore, we found that the distribution of transposable elements differs between the lines, with a higher load in monogamous lines. Our results suggest that complex interactions between purifying selection and sexual conflict are shaping the genome, particularly on chromosome 3 and the sex chromosome; sexual selection influences divergence across chromosomes but in a more complex way than proposed by simple 'purging' of deleterious loci.

Keywords: distribution of fitness effects; mutational load; sexual conflict; sexual selection.

Figure 1.

F_st of chromosomes 2, 3, 4 and X (split into XL and XR) between ‘Enhanced Polyandry’ and ‘Enforced Monogamy’ from generations 85 (blue) and 200 (red) of the sexual selection evolution experiment. Black horizontal bar shows F_st = 0.25. Individuals from each treatment and timepoint were sequenced and pooled into treatment. F_st analysis was carried out on 10 kilobase (kb) windows.

Figure 2.

Levels of π at chromosomes XL, XR and 3:22.5−23.5 Mb, i.e. regions of greatest genomic divergence. On the left-hand side, π in each region is displayed as boxplots for each replicate (side grey boxes) and treatment (Enforced Monogamy (M) (blue) and Elevated Polyandry (E) (red)) for generation 85 and 200. On the right-hand side are estimated marginal means of π for each timepoint and treatment averaged over replicates.

Figure 3.

(a) Total number of derived non-synonymous mutations across the entire genome for samples at generation 200 in ‘Enforced Monogamy’ (M) (blue) and ‘Elevated Polyandry’ (E) (red) treatments. Normality of data was confirmed for both treatments (M: W = 0.985, p = 0.970; E: W = 0.982, p = 0.803) and variance of dN was significantly lower in M than E lines (F_{1, 58} = 4.305, p = 0.043). (b) Mean number of derived non-synonymous mutations per coding sequence base pair (dN/CDS bp) for each major chromosome. Boxplots show range, median, interquartile range (IQR) and 1.5 × IQR whiskers.

Figure 4.

Number of identified transposable element (TE) sequences found in Enforced Monogamy (M, blue) and Enhanced Polyandry (E, red) lines. (a) Overall number of TE sequences found in each sequenced individual in each line. This is further broken down into (b) the number of TE sequences found in each chromosome from each sequenced individual and (c) the number of TE sequences identified from different TE families from each sequenced individual. Boxplots show range, median, interquartile range (IQR) and 1.5 × IQR whiskers. Asterisk between experimental evolution lines indicates p < 0.05.

Figure 5.

Distribution of deleterious fitness effects (DFE) for the Drosophila pseudoobscura evolved under ‘Enforced Monogamy’ (top) and ‘Elevated Polyandry’ (bottom) at each chromosome. Shown are the DFE at generation 85 (blue) and generation 200 (red). Bars show the mean proportion of derived allele effects and 95% confidence intervals (2 s.d. from the mean) at each respective discretized selection coefficient calculated from 500 bootstrapped site frequency spectra profiles. Discretized selection coefficient, with a lower value indicating a stronger deleterious effect, is shown on the x-axis. The y-axis is the percentage of derived allele effects at the corresponding discretized fitness group. Chromosome numbers and X arms are indicated within the grey boxes.

Figure 6.

Distribution of adaptive fitness effects (DFE) for the Drosophila pseudoobscura evolved under ‘Enforced Monogamy’ (top) and ‘Elevated Polyandry’ (bottom) at each chromosome. Shown are the DFE at generation 85 (blue) and generation 200 (red). Bars show the mean proportion of derived allele effects and 95% confidence intervals (2 s.d. from the mean) at each respective discretized selection coefficient calculated from 500 bootstrapped site frequency spectra profiles. Discretized selection coefficient, with a higher value indicating a more beneficial effect, is shown on the x-axis. The y-axis is the percentage of derived allele effects at the corresponding discretized fitness group. Chromosome numbers and X arms are indicated within the grey boxes.