Accumulating evidence indicates that some protein-coding genes have originated de novo from previously non-coding genomic sequences. However, the processes underlying de novo gene birth are still enigmatic. In particular, the appearance of a new functional protein seems highly improbable unless there is already a pool of neutrally evolving peptides that are translated at significant levels and that can at some point acquire new functions. Here, we use deep ribosome-profiling sequencing data, together with proteomics and single nucleotide polymorphism information, to search for these peptides. We find hundreds of open reading frames that are translated and that show no evolutionary conservation or selective constraints. These data suggest that the translation of these neutrally evolving peptides may be facilitated by the chance occurrence of open reading frames with a favourable codon composition. We conclude that the pervasive translation of the transcriptome provides plenty of material for the evolution of new functional proteins.

Cells express thousands of transcripts that show weak coding potential. Known as long non-coding RNAs (lncRNAs), they typically contain short open reading frames 5 (ORFs) having no homology with known proteins. Recent studies show that a significant proportion of lncRNAs are translated, challenging the view that they are non-coding. These results are based on selective sequencing of ribosome-protected fragments, or ribosome profiling. The present study used ribosome profiling data from eight mouse tissues and cell types, combined with ~330,000 synonymous and 10 non-synonymous single nucleotide variants, to dissect the patterns of purifying selection in proteins translated from lncRNAs. Using the three-nucleotide read periodicity that characterizes actively translated regions, we identified 832 mouse translated lncRNAs. Overall, they produced 1,489 different proteins, most of them smaller than 100 amino acids. Nearly half of the ORFs then showed sequence 15 conservation in rat and/or human transcripts, and many of them are likely to encode functional micropeptides, including the recently discovered Myoregulin. For lncRNAs not conserved in rats or humans, the ORF codon usage bias distinguished between two classes, one with particularly high coding scores and evidence of purifying selection, consistent with the presence of lineage-specific functional proteins, and a 20 second, larger, class of ORFs producing peptides with no significant purifying selection signatures. We obtained evidence that the translation of these lncRNAs depends on the chance occurrence of ORFs with a favorable codon composition. Some of these lncRNAs may be precursors of novel protein-coding genes, filling a gap in our current understanding of de novo gene birth.