Impact of GC content on de novo gene birth.

Noncoding regions in eukaryotes are extensively expressed and represent a significant source of novel microproteins, some of which become fixed as de novo genes. However, the structural properties of these unevolved products and the features driving their fixation remain poorly understood. Particularly, the influence of nucleotide composition (GC content) on their structural properties and evolutionary trajectories is still unclear. Here, we predict the foldability and sequence properties of millions of microproteins potentially encoded in the noncoding open reading frames (ORFs) of 3,379 eukaryotic genomes with GC contents ranging from 18% to 79%. Depending on GC content, these microproteins exhibit distinct structural properties, suggesting different cellular impacts if non-genic regions are pervasively expressed. Using phylostratigraphy, de novo gene search, and ancestral sequence reconstruction, we trace the evolution of several hundred de novo proteins across 22 organisms with varying GC contents. We show that de novo genes preferentially emerge from GC-rich ORFs with folding potential, revealing that the interplay between GC content and foldability - rooted in the structure of the genetic code - shapes the emergence of novel genes.