Subfunctionalization and Nonfunctionalization, Revisited

The idea of subfunctionalization (Alan Force and coworkers,  Genetics 151, 1531):  different, independent,  nearly neutral cis degenerative mutations within each duplicate gene following a tetraploidy should preserve the pair “forever” and explain why “too many” pairs were retained post-polyploidy.  However beautiful, subfunctionalization did not explain the majority of gene retention data following tandem versus whole genome duplications (Freeling, 2009, Ann Rev Plant Biol 60, 433).  Another  beautiful idea called “the gene balance hypothesis”  (E.g. Birchler and Veitia, 2007, The Plant Cell 19, 395)  trumped subfunctionalization as the one best explanation; removing only one duplicate tended to upset balance and was selected against.--  We now know that the subgenomes generated by allopolyploidy differ  (called “genome dominance”; Garsmeur, Schnable, J.C. et al. 2014, Mol Biol Evol 31, 448), thus complicating subfunctionalization theory.  Additionally, my lab’s new data that measures the deletion of cis-sites directly in post-polyploid pairs suggests that subfunctionalization does occur frequently enough to preserve pairs, but ONLY for genes with particularly complex  cis-regulatory regions. --   “Balanced gene drive” is one of the few hypothetical mechanisms that predictably drives increases in certain sorts of complexity as spandrels of polyploidy (Thomas and Freeling, 2006, Genome Research 16, 805).  Subfunctionalization by definition drives evolution nowhere.  This  re-evaluation of subfunctionalization is fundamental to evolutionary theory and rekindles a debate (Goldschmidt, 1953, Amer. Sci. 40, 84) that many think the long-deceased R. B. Goldschmidt lost definitively by 1950;  his signature evolutionary ideas are down, but not out, as will be explained.