Doug Axe: The difficulty to test evolution, Undeniable, page 27:
While I was doing the work that led to the 2000 JMB paper,1 a colleague of mine, Myriam Altamirano, was attempting to re-engineer a natural enzyme in order to make it perform the function of a different enzyme. Like many other scientists at the time, she was using a hybrid approach that combined aspects of design with aspects of evolution. In all of these projects, the idea was first to make informed guesses as to what parts of the original enzyme should be changed and how, and then, after implementing these changes, to use the standard laboratory version of evolution (mutate select repeat) to sort out any minor problems. Although this strategy could work in theory, the limitations have become increasingly apparent in the years since. Eleven years later, some of the leaders in the field conceded that “efforts to date to generate novel catalysts have primarily demonstrated that we are getting good at making bad enzymes. Making good enzymes will require a whole new level of insight, or new methodologies altogether.”2 The crux of the problem is that the evolutionary step at the end accomplishes so little that success rests almost entirely on the ability to make the right guesses in the first place. But, of course, if we knew how to do that, the evolutionary step would be largely superfluous. In other words, evolution seems to be an inadequate replacement for knowledge. Indeed, if our design intuition proves true, nothing is an adequate replacement for knowledge. Very good informed guesses, however, are tantamount to knowledge, and in this case Myriam’s guesses seemed to be that good. She found that her evolved engineered enzyme worked as well as the natural enzyme it was designed to imitate—a remarkable feat in a field where the term “success” usually had to be applied very generously. After writing up her results, Myriam submitted her paper for publication in the prestigious journal Nature around the time I met with Max Perutz. Her paper passed Nature’s peer review and appeared in February of 2000.
While I was doing the work that led to the 2000 JMB paper,1 a colleague of mine, Myriam Altamirano, was attempting to re-engineer a natural enzyme in order to make it perform the function of a different enzyme. Like many other scientists at the time, she was using a hybrid approach that combined aspects of design with aspects of evolution. In all of these projects, the idea was first to make informed guesses as to what parts of the original enzyme should be changed and how, and then, after implementing these changes, to use the standard laboratory version of evolution (mutate select repeat) to sort out any minor problems. Although this strategy could work in theory, the limitations have become increasingly apparent in the years since. Eleven years later, some of the leaders in the field conceded that “efforts to date to generate novel catalysts have primarily demonstrated that we are getting good at making bad enzymes. Making good enzymes will require a whole new level of insight, or new methodologies altogether.”2 The crux of the problem is that the evolutionary step at the end accomplishes so little that success rests almost entirely on the ability to make the right guesses in the first place. But, of course, if we knew how to do that, the evolutionary step would be largely superfluous. In other words, evolution seems to be an inadequate replacement for knowledge. Indeed, if our design intuition proves true, nothing is an adequate replacement for knowledge. Very good informed guesses, however, are tantamount to knowledge, and in this case Myriam’s guesses seemed to be that good. She found that her evolved engineered enzyme worked as well as the natural enzyme it was designed to imitate—a remarkable feat in a field where the term “success” usually had to be applied very generously. After writing up her results, Myriam submitted her paper for publication in the prestigious journal Nature around the time I met with Max Perutz. Her paper passed Nature’s peer review and appeared in February of 2000.
