By the way, here’s an article on the DFE for three different catalytic functions for a single enzyme.
Pay special attention to figure 4:
( a ) The DFE of missense and nonsense mutations for ACT (cranberry), PR (orange) and IB (cyan) growth selections. The dashed vertical line demarcates the wild-type fitness metric ( ζ =0.0). ( b ) DFE for beneficial mutations identified in the ACT, PR and IB selections, respectively. Overlain curves are best-fit exponential distributions estimated from the data.
Somehow this massive number of mutations of invisibly small fitness effect are absent, most mutations have significant effects tunable by selection, and the proportion of mutations that are beneficial range from about 22% to 5%.
Missense mutations were on average deleterious for the ACT and PR selections, with 75.8% and 74.2% of variants yielding at least 20% reduction in growth rate relative to wild-type, respectively. By contrast, only 45.4% fell below this threshold for the IB selection. Remarkably, 21.5% ( n =1,394) of missense mutations had above wild-type fitness metrics for the IB selection, with 483 (7.5%) variants having at least 10% increased growth rate ( ζ >0.15). There were appreciably less enhanced variants found in the ACT and PR selections, with 4.7% and 5.1% ( n =306 and 328) having fitness metrics above wild-type, respectively.
Wrap your head around those numbers. While most mutations are deleterious, the proportions are enormously removed from those assumed in Sanford’s work on GE, who is usually operating on the assumption that one in one million mutations are beneficial, with the rest being deleterious. Sanford will some times relax that assumption down to a range where 1 in 1000 mutations are beneficial, but considers basically any higher proportion to be physically unrealistic.
Price stated that proportions such as those measured(not drawn from analogies or assumed, measured in experiments) above are “self evidently false”. Well, I guess creationist intuitions about how “computers” work are of no real value in understanding the phenomenon of life, or the relationship between genetic mutations and the fitness effects of genotypes.
Notice something else there. The proportion of beneficial mutations is much higher for a non-canonical enzyme substrate the enzyme has not normally been selected on (fig4 a, blue color distribution). Which again supports the concept I have brought up multiple times before in GE discussions, which is the concept of diminishing returns epistasis, whereby the distribution of fitness effects of mutations changes as a function of fitness and gradually shifts the distribution towards more deleterious at higher fitness levels because the magnitude of effect and the pool of fitness-improving mutations gets closer to becoming exhausted for individual fitness-contributing functions, such as those of enzymes.