Those who follow the evolution intelligent/design are probably familiar with one accusation that we often face — that we do not do peer-reviewed research. In fact, we have published a number of peer-reviewed articles and one of them in particular, called “The Evolutionary Accessibility of New Enzyme Functions: A Case Study from the Biotin Pathway,” contributed to the argument we made in the new book published by Discovery Institute Press, Science and Human Origins.
That paper was critiqued last November by Paul McBride. He has now extended that critique in his review of our book, available here. McBride’s main complaint is that we picked an unnatural evolutionary transition to test. We chose to examine how hard it would be to get a modern-day enzyme to switch to the chemistry of a closely related modern-day enzyme, with very similar structures and catalytic mechanisms.
The reason for our choice was not ignorance. We knew that the enzymes we tested were modern, and that one was not the ancestor of the other. (They are, however, among the most structurally similar members of their family, and share many aspects of their reaction mechanism, but their chemistry itself is different.) We also knew that in order for a Darwinian process to generate the mechanistically and chemically diverse families of enzymes that are present in modern organisms, something like the functional conversion of one of these enzyme to the other must be possible. We reasoned that if these two enzymes could not be reconfigured through a gradual process of mutation and selection, then the Darwinian explanation of gene duplication and gradual divergence to new functions was called into question.
Our results indicated that a minimum of seven mutations would be required to convert or reconfigure one enzyme toward the other’s function. No one disputes that part of our research. What Paul McBride and others claim is that because we didn’t start from an “ancestral” enzyme, our results mean nothing. They say something like, “Of course transitions to new chemistries between modern enzymes are difficult. What you should have done is to reconstruct the ancestral form and use it as a starting point.”
Have you noticed the assumption underlying this critique? The assumption is that genuine conversions can be achieved only if you start from just the right ancestral protein. Why is that? Because conversions are hard.
McBride said as much in his post, tacitly acknowledging the legitimacy of our results, in the following quote:

Any biologist or biochemist could imagine useful new molecules in a given species that would aid their survival. Little imagination is needed, as many examples are found in other species. A simple example: an enzyme that breaks down cellulose into simple sugars would be immensely beneficial for virtually any heterotroph, yet such cellulases are only found in a handful of organisms, restricted to certain clades. Evolution is not a process that is capable of producing anything and everything, at all times in all species. It is, conversely, a greatly constrained process. A developmental biologist such as PZ Myers knows the minutiae of this constraint in particular models. Much of the process of evolution is guided by purifying selection – pruning those mutants that are at relative disadvantages to the general population – and most of the genomic change that does spread through populations is neutral and escapes selection altogether. Yes, transitioning between different enzyme functions is hard, but this is evidenced by it being relatively uncommon. In a broader sense, and to reiterate, many of the possible variations on life that we could imagine to exist do not exist. [Emphasis added.]

The problem then becomes, where did the diverse families of enzymes come from, if transitions are so hard, evolution is so constrained, and selection is so weak? Were the ur-proteins from which present families sprang so different from modern ones, so elastic that they could be easily molded to perform multiple functions? If so, how did they accomplish the specific necessary tasks for metabolism, transcription, and replication?
More than that, how did the proteins necessary for replication, transcription, translation, and metabolism arrive at all, if evolution is so constrained? Those processes are much more complicated that a cellulase enzyme. We have ribosomes, spliceosomes, photosynthesis, and respiration. We have hummingbirds and carnivorous plants and even cows who make use of cellulose-degrading symbionts. The things that have not arrived or arrived very rarely, like cellulases, seem trivial by comparison to the things we see around us.
Our results argue that only guided evolution, or intelligent design, can produce genuine innovations from a starting point of zero target activity. But McBride argues that we are the product of happenstance.

Evolution is a process without teleology and long-term targets or goals. Natural selection can provide relatively short-term direction along ‘local’ fitness gradients, which may be helpful or unhelpful; escapes from selection are also predicted to be important in many evolutionary paths. This could be a problem to neofunctionalisation where teleology is invoked, except that no particular enzymes were ever mandated to evolve. Life would have been different if particular enzymes that do exist had not arisen, but some other suite of enzymes would undoubtedly exist instead had the dice been rolled differently. Life would very much go on. It is a fairly safe conjecture that only a small number out of all the possible enzymes exist, and many of these exist only in small clades in the tree of life. [Emphasis added]

McBride argues against teleology and opts for chance. He is more sanguine than I about a new “suite of enzymes” evolving, given the apparent difficulty with which they evolve. Life is inherently teleological, and the needs of an organism cannot be met by whatever happens to show up. I would say, rather, that his faith in the unending creativity of evolution, in spite of the limitations of natural selection, the rarity of paths, and the functional needs of organisms, is itself a form of religion.
This is an interesting turn in evolutionary thinking. People have been saying for years, “Of course evolution isn’t random, it’s directed by natural selection. It’s not chance, it’s chance and necessity.” But in recent years the rhetoric has changed. Now evolution is constrained. Not all options are open, and natural selection is not the major player, it’s the happenstance of genetic drift that drives change. But somehow it all happens anyway, and evolution gets the credit.
All around us we see marvelous examples of successful, even optimal design. If evolution is constrained to just a few paths, and you have to start with the right ancestral form to get anywhere, and fixation of useful new traits happens by accident, how did anything ever happen at all? Were the paths of adaptation “preordained”? Paul’s choice of words, not mine. If there are only a few ways to solve any problem set by the needs of the organism because transitions are hard, then either the deck was stacked in our favor, or the process was guided, or we are incredibly lucky. That might be called preordained, I suppose.