Comparative evolutionary genomics revealed that novel protein coding genes can emerge
randomly from non-coding DNA. While most of the myriads of transcripts which continuously
emerge vanish rapidly, some attain regulatory regions, become translated and survive.
More surprisingly, sequence properties of de novo proteins are almost indistinguishable
from randomly obtained sequences, yet de novo proteins may gain functions and integrate
into eukaryotic cellular networks quite easily. We here discuss current knowledge on de novo
proteins, their structures, functions and evolution. Since the existence of de novo proteins
seems at odds with decade-long attempts to construct proteins with novel structures and
functions from scratch, we suggest that a better understanding of de novo protein evolution
may fuel new strategies for protein design.