Abstract
Genomes can encode a variety of proteins with unrelated architectures and activities. It is known that protein-coding genes of de novo origin have significantly contributed to this diversity. However, the molecular mechanisms and evolutionary processes behind these originations are still poorly understood. Here we show that the last 102 codons of a novel gene, Noble, assembled directly from non-coding DNA following an intronic deletion that induced alternative intron retention at the Drosophila melanogaster Rieske Iron Sulphur Protein (RFeSP) locus. A systematic analysis of the evolutionary processes behind the origin of Noble showed that its emergence was strongly biased by natural selection on and around the RFeSP locus. Noble mRNA is shown to encode a bona fide protein that lacks an iron sulphur domain and localizes to mitochondria. Together, these results demonstrate the generation of a novel protein at a naturally selected site.