Here’s a paper I read recently that I think presents some neat ideas (though it’s not without its faults):
Nature. 2012 Jul 19;487(7407):370-4. doi: 10.1038/nature11184.
Proto-genes and de novo gene birth.
Carvunis AR, Rolland T, Wapinski I, Calderwood MA, Yildirim MA, Simonis N, Charloteaux B, Hidalgo CA, Barbette J, Santhanam B, Brar GA, Weissman JS, Regev A, Thierry-Mieg N, Cusick ME, Vidal M.
Essentially they point out the fact that lots of open reading frames in non-genic sequences are translated in yeast, and then explore their potential to become “proto-genes” which then leads to de novo gene birth. Some basic points/things that stuck out to me:
- They believe de novo gene birth from these protogenes may be more prevalent than gene birth from the re-organization of existing genes e.g. through gene duplication and subsequent sub/neo functionalization.
- Along the way they contrast proto-genes and gene birth with pseudo-genes and gene death, pointing out that the formation of pseudo-genes is reversible (i.e. a gene can “come back from the dead” – I think this is pretty hotly contested), but the formation of proto-genes is not reversible (i.e. a gene cannot revert to being a proto-gene)
- They claim the key is in the translation of these non-genic sequences, and that this provides material for selection to act on.
- They propose a model for identifying genes on a continuum, rather than having two categories of “annotated” and “unannotated” genes.
Any thoughts? I really like the way they presented their analysis – they came up with a model for how proto-genes might evolve, made predictions based on that model, and then verified their predictions (to some extent). The only thing is, I think some of the conclusions they make about the adaptive potential of these translated ORFs and their predicted structure aren’t super clear from the data. Regardless, it’s a pretty enjoyable read.