Genome-wide analysis of retrogene polymorphisms in Drosophila melanogaster

Abstract

Gene duplication via retrotransposition has been shown to be an important mechanism in evolution, affecting gene dosage and allowing for the acquisition of new gene functions. Although fixed retrotransposed genes have been found in a variety of species, very little effort has been made to identify retrogene polymorphisms. Here, we examine 37 Illumina-sequenced North American Drosophila melanogaster inbred lines and present the first ever data set and analysis of polymorphic retrogenes in Drosophila. We show that this type of polymorphism is quite common, with any two gametes in the North American population differing in the presence or absence of six retrogenes, accounting for ~13% of gene copy-number heterozygosity. These retrogenes were identified by a straightforward method that can be applied using any type of DNA sequencing data. We also use a variant of this method to conduct a genome-wide scan for intron presence/absence polymorphisms, and show that any two chromosomes in the population likely differ in the presence of multiple introns. We show that these polymorphisms are all in fact deletions rather than intron gain events present in the reference genome. Finally, by leveraging the known location of the parental genes that give rise to the retrogene polymorphisms, we provide direct evidence that natural selection is responsible for the excess of fixations of retrogenes moving off of the X chromosome in Drosophila. Further efforts to identify retrogene and intron presence/absence polymorphisms will undoubtedly improve our understanding of the evolution of gene copy number and gene structure.

Figure 1.

Mapping reads from a genome containing a polymorphic retrocopy to a reference genome. The black line at the top represents a chromosome in a sample genome. Gray boxes represent exons within a gene, with the spaces in between representing introns. The gray boxes on the right with no introns in between them represent a retrogene derived from a parental gene (located downstream in this example). The short bars appearing above the two gene copies represent reads derived from the sample chromosome. The black line at the bottom represents the same chromosome in the reference genome to which these reads are mapped. Note that reads derived from the parental copy of the gene are mapped to the proper location, while reads from the retrocopy (light gray) are mapped to the exons of the parental copy, resulting in elevated read-depth. Also note that the reads crossing exon–exon boundaries in the retrocopy (dark gray) are not mapped to the reference genome. Our method to detect retroCNVs involves finding these reads by searching all unmapped reads against a database of exon–exon junctions.

Figure 2.

Mapping reads from a genome containing a polymorphic intron deletion to a reference genome. As in Figure 1, the black line at the top represents a chromosome in a sample genome, the bottom line represents the chromosome in the reference genome, gray boxes represent exons, and short bars represent reads. Note that no reads are mapped to the intron that is deleted in the sample chromosome but present in the reference genome. Also note that the reads crossing the single exon–exon boundary in the sample chromosome are not mapped to the reference genome.

Figure 3.

Derived allele frequency spectrum of retroCNVs. The derived allele frequency is given as the number of sequenced lines containing a retrocopy.

Figure 4.

Derived allele frequency spectrum of intron deletions. The derived allele frequency is given as the number of sequenced lines lacking the intron.