Recombination yet inefficient selection along the Drosophila melanogaster subgroup's fourth chromosome

Abstract

A central goal of evolutionary genetics is an understanding of the forces responsible for the observed variation, both within and between species. Theoretical and empirical work have demonstrated that genetic recombination contributes to this variation by breaking down linkage between nucleotide sites, thus allowing them to behave independently and for selective forces to act efficiently on them. The Drosophila fourth chromosome, which is believed to experience no-or very low-rates of recombination has been an important model for investigating these effects. Despite previous efforts, central questions regarding the extent of recombination and the predominant modes of selection acting on it remain open. In order to more comprehensively test hypotheses regarding recombination and its potential influence on selection along the fourth chromosome, we have resequenced regions from most of its genes from Drosophila melanogaster, D. simulans, and D. yakuba. These data, along with available outgroup sequence, demonstrate that recombination is low but significantly greater than zero for the three species. Despite there being recombination, there is strong evidence that its frequency is low enough to have rendered selection relatively inefficient. The signatures of relaxed constraint can be detected at both the level of polymorphism and divergence.

FIG. 1.

Upper panel: Regression of the recombination estimate r² over SNP distance. Lower panel: Empirical distribution of r² values for permuted r² versus distance samples. Dark vertical line indicates the true estimate.

FIG. 2.

Jackknife resampling of the composite likelihood estimates for the population recombination rate, ρ, and the ratio of conversion to crossing over, f. To facilitate comparisons across species with different effective population sizes, ρ has been scaled by the respective species silent diversity, θ_s. (A) Pseudovalues for the joint distributions of ρ and f. (B) Jackknife means and 95% CIs for ρ. (C) Jackknife means and 95% CIs for f.

FIG. 3.

Comparisons of nucleotide diversity estimates. (A) Boxplot of Wattersons θ (W) and π (Pi) for the unpartitioned loci between the three species. (B–D) Boxplots of replacement (rep) and silent (sil) diversity. For each species, there is significantly greater silent diversity. (E–F) Comparisons of the replacement/silent diversity between the three species. Vertical bars represent the 95% bootstrap CI.

FIG. 4.

Box plots summarizing divergence comparisons. For all comparisons the ratio dN/dS is also broken down to display dN and dS alone. Top panels are between species comparisons of the fourth chromosome. Bottom panels are within species comparisons of the fourth chromosome (4th) to other regions of the genome experiencing a range in the amount of recombination (auto = nonheterochromatic autosomal loci, X = X chromosome loci, hetero = heterochromatic autosomal loci), as defined in Begun et al. (2007).

FIG. 5.

Box plots comparing estimates of codon bias (effective number of codons, ENC) from chromosomes 2 and 3 to estimates from the fourth chromosome.

FIG. 6.

Estimates of α, the proportion of amino acid substitutions driven by positive selection. The method for the two estimates is provided above each column (see Materials and Methods). Two values for each estimate are provided, each one using a different outgroup species (Dy, D. yakuba; Dm, D. melanogaster; and Ds, D. simulans). Vertical bars are 95% boostrap CIs.