Eye Evolution: The Waiting Is the Hardest Part
Without calling it a series, I've written several articles recently that followed a logical path. In the first, I described the distinction between incremental innovation and radical innovation. I also outlined the commonalities and differences between intelligent design and theistic evolution (TE) as approaches to biology. In a follow-up, I applied the concepts from the first article to the proposed evolution of the vertebrate eye, demonstrating that it could not have occurred without intelligent direction. That's mainly because the majority of steps required for the addition of a lens are disadvantageous in isolation, so selective pressures would have operated in opposition to the evolutionary process.
Let's now consider the challenge of waiting times -- the minimum time required for hypothesized evolutionary transformations, such as the development of the camera eye, to occur through undirected processes. Even if the selective pressures were favorable, the required timescales are far longer for sufficient numbers of coordinated mutations to accumulate than the maximum time available, as determined by the fossil record. Of special interest is the proposed cooption of crystallin proteins, which give the lens its refractive properties. Seemingly, one of the easiest evolutionary steps should be producing these proteins in the lens, for some of them are already used for other purposes. The main hurdle would simply be altering the regulatory regions of the first borrowed crystallin gene and other related genes, so they bind to the correct set of transcription factors (TFs). The lens protein could then be overexpressed in the fiber cells in sufficient quantities at the right time in development.
However, the cooption process is far more challenging than it might at first appear. It requires specific regulatory regions to bind to well over four new transcription factors. This alteration would involve numerous mutations creating over four corresponding DNA binding sites known as transcription factor binding sites (TFBS). As I mentioned in the previous article, the earliest lens should have closely resembled lenses of vertebrates today, so this lower estimate is almost certainly accurate.
A typical binding site involved in lens construction consists of a DNA sequence ranging from roughly 6 (e.g., SOX2) to 15 (e.g., Pax6) base pairs, so four TFBS would likely correspond to over 30 base pairs. One could think of these DNA sequences like the launch codes to a missile; they must be correct before the protein can be properly manufactured. The lower bound of 30 base pairs can be divided by a factor of 3 to compensate for sequence redundancies, flexibility in where in the DNA sequences start, and the fact that roughly one quarter of the bases would be correct purely by chance. This extremely conservative estimate indicates that over 10 mutations would be required to generate a proper sequence. All but the final mutation would be neutral.
We can now calculate the likelihood of sufficient mutations occurring in 10 million generations. The mutation rate for a specific base par is typically estimated for complex animals to correspond to a probability around 1 in 100 million. The chance of a mutation occurring in 10 million generations is then 1 in 10. Therefore, the chance of 10 coordinated mutations appearing on the same DNA strand works out to much less than 1 in 10 billion. No potential precursor to a vertebrate with a lens would have had an effective population large enough to acquire the needed mutations. For comparison, the effective population size estimate used for Drosophila melanogaster can be in the low millions. If the generation time were even as low as one year, a crystallin could not be coopted even in 10 million years, which is the time required for the appearance of most known phyla in the Cambrian explosion.
Moreover, this step is only one of hundreds required to produce a lens. Researchers have identified
Biologists have claimed to produce viable scenarios for the evolution of several other complex systems. What all these stories share is that they ignore crucial details and lack careful analysis of feasibility. When we examine these issues in detail, the stories collapse for the same reasons that the one about the eye does: First, the selective pressures oppose transitions between key proposed stages. Second, the required timescales are vastly longer than what is available.
For biologists, rigorously evaluating evolutionary narratives has become fully possible only in the past several decades due to advances in molecular and developmental biology. Meanwhile, with breakthroughs in computer engineering, information theory, and nanotechnology, parallels between biological and human engineered systems are increasingly evident. These developments are making the intelligent design framework essential for scientific advancement. They also create new opportunities for ID proponents and theistic evolutionists to collaborate.
Proponents of TE want to push materialistic explanations for biological systems as far as possible, as science demands. ID advocates would not disagree with them on that. No one wants to trigger the design filter prematurely. So theistic evolutionists should join us in considering what the modern evolutionary synthesis with its auxiliary hypotheses, such as niche construction and epigenetic inheritance, can explain. We should all continue to examine how insights from evolution may benefit research on cancer, in epidemiology, and other fields.
ID researchers, meanwhile, can examine the limits of purely materialistic processes, and we invite theistic evolutionists to do likewise These combined efforts will help to define in greater detail what Michael Behe calls the edge of evolution. This understanding would also help advance research on cancer treatments, antibiotic protocols, and more. At the same time, ID proponents can help identify how principles and insights from engineering may advance biological research and related applications.
Many theistic evolutionists recognize that the appearance of design is real (but then, so does Richard Dawkins). This insight, at least, should inform their research. In contrast, anti-theistic evolutionists are biased against recognizing the benefits of design thinking. As a result, in studying life they have stumbled upon close parallels to human engineering, which, however, they recognized only begrudgingly. On the other hand, ID expects these parallel and is unsurprised to find them. A classic example is how researchers, misled by evolutionary thinking, dismissed a large portion of the human genome as "junk" DNA instead of anticipating that it would function as a genomic operating system.
TE researchers do not need to immediately agree with ID researchers on whether any particular feature of life is the result of primary design or secondary causes. They can still work together to best serve the cause of genuine science, and I hope they will do so more in the future.