ABSTRACT Taxonomically restricted genes are increasingly understood to play major roles in evolution. However, a significant body of work has taken issue with the notion of widespread “novel” genes and argued that such genes have homologs in distant clades that can be found with either sufficiently powerful alignment techniques or by using synteny to find their ancestral sequences. Here, we argue that such work is misguided. Moreover, we argue that the whole notion of genetic assignment of function (and annotation) based on historical origin violates the levels of analysis distinction between origin and current utility. The evolutionary history of a gene is so often not reflective of its current utility that naming genes based on the function of their homologs is bad practice. This is nowhere more apparent than in the case of genes that have changed so radically from their ancestors that they bear no similarity to them at the sequence or protein folding levels. We coin the term, overwriting, for this process in which selection creates novel genes by completely changing the coding sequence of a gene in a manner that does not conserve function , and argue for the general importance of this mechanism.
Brian R. Johnson (Tue,) studied this question.