Subfunctionalization versus neofunctionalization after whole-genome duplication

The question of what is the predominant evolutionary fate of genes after duplication events has been hotly debated for decades. Two recently published articles in Nature (Lien et al.) and Nature Genetics (Braasch et al.) investigated the regulatory fate of gene duplicates after the salmonid-specific (Ss4R) and teleost specific (Ts3R) whole genome duplication (WGD) events, respectively. Both studies relied on tissue expression atlases for estimating regulatory divergence and used closely related unduplicated sister taxa (i.e. Northern pike and the spotted gar, respectively) as proxies for the ancestral expression state. Surprisingly, the two studies reach very different conclusions about the evolutionary mechanisms impacting gene expression after WGD. Braasch et al. concluded that the expression divergence was consistent with partitioning of tissue regulation between duplicates (subfunctionalization), while Lien et al. concluded that most divergence in tissue regulation were consistent with one copy maintaining ancestral tissue regulation while the other having diverged (in line with neofunctionalization). Here we show that this striking discrepancy in the conclusions of the two studies is a consequence of the data analysis approaches used, and is not related to underlying differences in the data.