We continue to make our way through Christoph Adami’s book, The Evolution of Biological Information. Currently, we are going through the chapter on “Experiments in Evolution.” While writing about his own work on the Avida system, he discusses the evolution of the eye, attempting to present Avida as giving insight in the ability of Darwinian evolution to account for complex traits like the vertebrate eye.
Adami presents the eye as “a very instructive example of a complex trait that has gradually evolved.” He says that it evolves by the gradual addition of parts to an existing functional system. Each part is constructed by reusing and repurposing existing genes. He states:
To wit, people are often puzzled about how it is possible that a protein such as crystallin can evolve.
He goes on to explain that at least for the family of alpha-crystallin proteins, they would have already existed as chaperones, proteins that help other proteins fold. Thus, the origin of crystallin is well explained.
Or Is It?
The first problem is that it does not explain the origin of crystallin; rather, it depends on crystallin having already evolved for another purpose. How did it evolve in the first place as a chaperone? If it is a problem for it to have evolved in place as a lens protein, why does relocating its evolution to a chaperone protein change anything? The same challenge to evolving a novel protein exists there.
Even if crystallin has already evolved, what does repurposing it for the lens entail? At minimum, this would require precise modifications to gene regulation. It would also likely require modifications to the protein itself to make it suitable for the new role.
But There Is a Bigger Problem
Adami has slipped from the difficult problem of explaining the intricate complexity of the vertebrate eye to the relatively simple problem of explaining the origin of a single protein. The eye poses a challenge to evolutionary explanations because many interacting parts must come together. To focus on explaining a single protein is to leave everything that makes it difficult to focus on one part that you can construe as easy.
Adami goes through the historical process of how a simple nand function evolves in Avida. Avida uses English letters in its genomes, and a minimal version of this function can be coded with the letters: qcqpq. This sequence became functional when the last letter, in this case the “p,” was in place. This only worked because other letters, notably the three “q”s, were already in place. How did those letters come to be in place? They were simply neutral mutations, having no effect either way on fitness.
Adami says, “it is hard not to see the parallels to other evolutionary innovations, for example the co-option of the existing crystallin protein in the eye.” I must confess that I find the parallels hard to see. In the crystallin scenario, an existing functional protein is reused in a new role. In the Avida scenario, a random combination of mutations with no function of their own come together to produce a new function. Those are two completely different scenarios.
One might ask, why does it work for Avida to construct functions based on random combinations of mutations like this? Avida is deliberately designed to ensure that many random sequences have functionality. Furthermore, it utilizes a mutation rate much higher than biological systems allow it to try many more mutation combinations than would be realistic in a biological setting. In that case, it is feasible to assemble these short combinations to produce functionality.
Very Easy to Adapt
But one could claim that when Avida evolves more complex functions, like the EQU, it uses something like co-option because it adapts a functional system computing some less complex function into a more complex function. What is happening is that Avida is designed so that all functions are very similar to each other. It is thus very easy to adapt one function into another and thus attain more complex functionality.
Adami’s attempt to put Avida to service in making eye evolution seems plausible. But he first slips from the complexities of eye evolution to explaining a single protein. Even that fails, because he assumes that the protein must exist elsewhere and can be trivially adapted to its new role. His proposed parallel does not fit and even if we look at the moderately more complex function in Avida, their ease of evolution is due not to the power of the evolutionary process, but to the particulars of the design of the Avida system.