Population genomics reveals mechanisms and dynamics of de novo proto-gene emergence in Drosophila melanogaster

Scientific AbstractNovel genes are essential for evolutionary innovations and differ substantially even between closely related species. Recently, multiple studies across many taxa have suggested that some novel genes arise de novo, i.e. from previously non-coding DNA. In order to characterise the underlying mutations that allowed de novo gene emergence and their order of occurrence, homologous regions must be detected within non-coding sequences in closely related sister genomes. So far, most studies do not detect non-coding homologs of de novo genes due to inconsistent data and long evolutionary distances separating genomes. Here we overcome these issues by searching for proto-genes, the not-yet fixed precursors of de novo genes that emerged within a single species. We sequenced and assembled genomes with long-read technology and the corresponding transcriptomes from inbred lines of Drosophila melanogaster, derived from seven geographically diverse populations. We found line-specific proto-genes in abundance but few proto-genes shared by lines, suggesting a rapid turnover. Gain and loss of transcription is more frequent than the creation of Open Reading Frames (ORFs), e.g. by forming new START- and STOP-codons. Consequently, the gain of ORFs becomes rate limiting and is frequently the initial step in proto-gene emergence. Furthermore, Transposable Elements (TEs) are major drivers for intra genomic duplications of proto-genes, yet TE insertions are less important for the emergence of proto-genes. However, highly mutable genomic regions around TEs provide new features that enable gene birth. In conclusion, proto-genes have a high birth-death rate, are rapidly purged, but surviving proto-genes spread neutrally through populations and within genomes.