Estimation of fine-scale recombination intensity variation in the white-echinus interval of D. melanogaster

Abstract

Accurate assessment of local recombination rate variation is crucial for understanding the recombination process and for determining the impact of natural selection on linked sites. In Drosophila, local recombination intensity has been estimated primarily by statistical approaches, by estimating the local slope of the relationship between the physical and genetic maps. However, these estimates are limited in resolution and, as a result, the physical scale at which recombination intensity varies in Drosophila is largely unknown. Although there is some evidence suggesting as much as a 40-fold variation in crossover rate at a local scale in D. pseudoobscura, little is known about the fine-scale structure of recombination rate variation in D. melanogaster. Here we experimentally examine the fine-scale distribution of crossover events in a 1.2-Mb region on the D. melanogaster X chromosome using a classic genetic mapping approach. Our results show that crossover frequency is significantly heterogeneous within this region, varying approximately 3.5-fold. Simulations suggest that this degree of heterogeneity is sufficient to affect levels of standing nucleotide diversity, although the magnitude of this effect is small. We recover no statistical association between empirical estimates of nucleotide diversity and recombination intensity, which is likely due to the limited number of loci sampled in our population genetic data set. However, codon bias is significantly negatively correlated with fine-scale recombination intensity estimates, as expected. Our results shed light on the relevant physical scale to consider in evolutionary analyses relating to recombination rate and highlight the motivations to increase the resolution of the recombination map in Drosophila.

Figure 1

Schematic representation of our two-step crossing scheme. Boxed F2 male progeny correspond to the two recombinant genotypes identified by our screen; these males contain a single crossover in the region between white and echinus.

Figure 2

Recombination intensity (in cM/Mbp) as estimated by crossover frequency per physical distance. Error bars correspond to the 95% confidence interval. The intervals are defined by eight SNP markers in combination with the two visible phenotypic markers (white and echinus).

Figure 3

Representative simulation results. Boxplots correspond to distributions of nucleotide diversity as a function of recombination rate under different parameter combinations, with the selection parameter ranging from 0.1 to 10 and the probability of a novel mutation being advantageous ranging from 0.05 to 0.15. For these plots, the probability of a novel mutation being deleterious is constant at 0.5. The range of the population recombination parameter rho is consistent across plots, ranging from 0.01 to 1, and the leftmost two values in each plot (0.01 and 0.05) correspond roughly to the range of recombination intensity captured by our empirical study. The notch in each box corresponds to the median value, and the lower and upper edges of the box correspond to the 25^th and 75^th percentile, respectively. Whiskers extend to the most extreme datapoint which is no more than 1.5 times the interquartile range from the box.

Figure 4

Scatterplot of recombination intensity and nucleotide diversity (normalized by pairwise divergence with D. simulans) African and non-African populations.

Figure 5

Scatterplot of recombination intensity and a) codon bias, as measured by the frequency of optimal codons (FOP) and b) intronic GC content.