Genomic analysis of P elements in natural populations of Drosophila melanogaster

Abstract

The Drosophila melanogaster P transposable element provides one of the best cases of horizontal transfer of a mobile DNA sequence in eukaryotes. Invasion of natural populations by the P element has led to a syndrome of phenotypes known as P-M hybrid dysgenesis that emerges when strains differing in their P element composition mate and produce offspring. Despite extensive research on many aspects of P element biology, many questions remain about the genomic basis of variation in P-M dysgenesis phenotypes in natural populations. Here we compare gonadal dysgenesis phenotypes and genomic P element predictions for isofemale strains obtained from three worldwide populations of D. melanogaster to illuminate the molecular basis of natural variation in cytotype status. We show that the number of predicted P element insertions in genome sequences from isofemale strains is highly correlated across different bioinformatics methods, but the absolute number of insertions per strain is sensitive to method and filtering strategies. Regardless of method used, we find that the number of euchromatic P element insertions predicted per strain varies significantly across populations, with strains from a North American population having fewer P element insertions than strains from populations sampled in Europe or Africa. Despite these geographic differences, numbers of euchromatic P element insertions are not strongly correlated with the degree of gonadal dysgenesis exhibited by an isofemale strain. Thus, variation in P element insertion numbers across different populations does not necessarily lead to corresponding geographic differences in gonadal dysgenesis phenotypes. Additionally, we show that pool-seq samples can uncover population differences in the number of P element insertions observed from isofemale lines, but that efforts to rigorously detect differences in the number of P elements across populations using pool-seq data must properly control for read depth per strain. Our work supports the view that euchromatic P element copy number is not sufficient to explain variation in gonadal dysgenesis across strains of D. melanogaster, and informs future efforts to decode the genomic basis of geographic and temporal differences in P element induced phenotypes.