Darwin's Doubt: The Explosive Origin of Animal Life and the Case for Intelligent Design

by Stephen C. Meyer

Moreover, these scenarios not only assume unexplained preexisting sources of biological information, they do so without explaining or even attempting to explain how any of the mechanisms they envision could have solved the combinatorial search problem described in Chapters 9 and 10.

beg important questions about the origin of genetic information.

Each of these six mutational mechanisms presupposes preexisting modules of specified genetic information.

Some of these mutational mechanisms also depend upon sophisticated preexistent molecular machines such as the reverse transcriptase enzyme used in retropositioning or other complex cellular machinery involved in DNA replication. Since building these machines requires other sources of genetic information, scenarios that presuppose the availability of such molecular machines to assist in the cutting, splicing, or positioning of modular sections of genetic information clearly beg the question. Overall, what evolutionary biologists have in mind is something like trying to produce a new book by copying the pages of an existing book (gene duplication, lateral gene transfer, and transfer of mobile genetic elements), rearranging blocks of text on each page (exon shuffling, retropositioning, and gene fusion), making random spelling changes to words in each block of text (point mutations), and then randomly rearranging the new pages. Clearly, such random rearrangements and changes will have no realistic chance of generating a literary masterpiece, let alone a coherent read. That is to say, these processes will not likely generate specificity of arrangement and sequence and, therefore, do not solve the combinatorial search problem. In any case, all such scenarios also beg the question. There is a big difference between shuffling and slightly altering preexisting sequence-specific modules of functional information and explaining how those modules came to possess information-rich seque...

“60 new protein-coding genes that originated de novo on the human lineage since divergence from the chimpanzee,”25 a finding that was called “a lot higher than a previous, admittedly conservative, estimate.”

It investigated the de novo origin of genes and acknowledged, “The emergence of complete, functional genes—with promoters, open reading frames (ORFs), and functional proteins—from ‘junk’ DNA would seem highly improbable, almost like the elusive transmutation of lead into gold that was sought by medieval alchemists.”27 Nonetheless, the article asserted without saying how that: “evolution by natural selection can forge completely new functional elements from apparently nonfunctional DNA—the process by which molecular evolution turns lead into gold.”28

After one study of fruit flies reported that “as many as ~12% of newly emerged genes in the Drosophila melanogaster subgroup may have arisen de novo from noncoding DNA,”29 the author went on to acknowledge that invoking this “mechanism” poses a severe problem for evolutionary theory, since it doesn’t really explain the origin of any of its “nontrivial requirements for functionality.”

Thus, ultimately, the scenarios featured in Long’s review essay do not explain the origin of the specified information in either genes or sections of genes. That would require a cause capable of solving the combinatorial inflation problem discussed in the previous chapters. But none of the scenarios discussed in Long’s article even addresses this problem, let alone demonstrates the mathematical plausibility of the mechanisms they cite. Yet, Gishlick, Matzke, and Elsberry originally cited Long as a definitive refutation of my article—the one in which I argued that the rarity of genes and proteins in sequence space cast doubt on the power of selection and mutations to generate novel genetic information. Professor Miller, in his testimony at the celebrated Dover trial, even convinced a federal judge to affirm that Long had succeeded in demonstrating how genetic information originates in a celebrated legal ruling. Clearly, one cannot solve a problem or refute an argument by failing to address it.31

There is a second and closely related difficulty associated with the scenarios cited by Long. Typically, they do not even try to explain the origin of new protein folds, and few of them analyze genes different enough from each other that their protein products could, even conceivably, exemplify different folds.

Thus, at the very least, the exon-shuffling hypothesis presupposes the prior existence of a significant amount of genetic information—enough information to build an independent protein domain or fold. As such, it fails as an explanation for the origin of protein folds and the information necessary to produce them.

Thus, this version of the exon-shuffling hypothesis lacks credibility because it incorrectly assumes that shapeless protein fragments can be mixed and matched in a modular fashion to form new stable, functional protein folds.

It still presupposes unexplained functional information—in particular, the information necessary to specify, not just the smaller fragments, but also the information required to arrange these smaller units into stable folds, and ultimately functional proteins.

The need for extreme specificity in the sequential arrangements of amino acids, discussed in the previous chapter, means that the overwhelming majority of amino-acid sequences in units of secondary structure will not result in stable folds as these units of structure come into contact with each other. As discussed in Chapter 10, the extreme rarity of functional proteins (with stable folds) in sequence space ensures that the probability of finding a correct fold-stabilizing sequence will be astonishingly small. For this reason, even skilled protein scientists have struggled to design sequences that will produce stable protein folds.42 Almost invariably the units of secondary structure that they attempt to combine or otherwise place into stable composite structures will not fold because of the interactions of their amino-acid side chains.

43 As molecular biologist Ann Gauger explains, “Thus, [alpha] helices and [beta] sheets are sequence-dependent structural elements within protein folds. You can’t swap them around like Lego bricks.”44

Generating specific sequences that will fold into stable structures, whether in the lab or during the history of life, requires solving the combinatorial inflation problem.

All this requires searching for a functional needle in a vast haystack of combinatorial possibilities. Recall that Douglas Axe estimated the ratio of needles (functional sequences) to strands of straw in the haystack (nonfunctional sequences) to be 1 to 1077 for sequences of modest-length (150 amino acids).

The question is not whether the combinatorial search problem necessary to produce stable protein folds has ever been solved, but whether a neo-Darwinian mechanism relying on random mutations (in this case random shuffling of exons) provides a plausible explanation for how it might have been solved.

The exon-shuffling hypothesis ignores the need for side-chain specificity, though the need for such specificity has repeatedly defeated attempts in the laboratory to build new proteins from units of secondary structure in the manner required. But advocates of exon shuffling make no attempt to show how random rearrangements of protein domains—whether the domains are conceived of as fragments of a fold or whole folds—would solve the combinatorial problem. Nor do they challenge Sauer’s or Axe’s experimentally derived quantitative estimates of the rarity of functional genes or proteins. They do not challenge the probability calculations based on these estimates. And they do not show that a mechanism exists that can search amino-acid sequence spaces more effectively or efficiently than random mutation and selection. Nor do they demonstrate the efficacy of exon shuffling in a model system in the laboratory. Instead, basic considerations of protein structure imply the implausibility of exon shuffling as a means of generating the genetic information necessary to produce a new protein fold.45 So, in the end, with few words and with apparent confidence, advocates of exon shuffling simply assert, as the Long paper does, that “exon shuffling often recombines sequences that encode various protein domains to create mosaic proteins.”

The assertion of Long and his colleagues about exon shuffling, like many other statements about postulated mutational mechanisms, blurs the distinction between theory and evidence. Despite the authoritative tone of such statements, evolutionary biologists rarely directly observe the mutational processes they envision. Instead, they see patterns of similarities and differences in genes and then attribute them to the processes they postulate. Yet the papers that Long cites offer neither mathematical demonstration, nor experimental evidence, of the power of these mechanisms to produce significant gains in biological information.

These vague narratives resemble nothing so much as the naming games of scholastic philosophers in the Middle Ages. Why does opium put people to sleep? Because it has a “dormative” virtue. What causes new genes to evolve so rapidly? Their “hypermutability” or perhaps their ability to undergo “rapid, adaptive evolution.” How do we explain the origin of two similar genes in two separate, but otherwise widely disparate lineages? Convergent evolution, of course. What is convergent evolution? The presence of two similar genes in two separate, but otherwise widely disparate lineages. How does convergent evolution occur, given the improbability of finding even one functional gene in sequence space, let alone the same gene arising twice independently? No one knows exactly, but perhaps it was a “fortuitous juxtaposition of suitable sequences,” or “positive selection,” or “de novo origination.” Need to explain two similar genes in more closely related lineages? Try “gene duplication,” or “chimerical gene fusion,” or “retropositioning,” or “extensive refashioning of the genome,” or some other scientific-sounding combination of words.

So despite the official pronouncement of a federal judge and claims of extensive “scientific literature documenting the origin of new genes,” evolutionary biologists have not demonstrated how new genetic information arises, at least not in amounts sufficient to build protein folds, the crucial units of biological innovation. Biologists have not solved the problem of combinatorial inflation or refuted the precise quantitative argument against the creative power of the selection and mutation mechanism presented in the previous chapter (or in my 2004 article). Nor has anyone provided a compelling refutation of Douglas Axe’s assessment of the rarity of genes and proteins on which that argument is based.

According to neo-Darwinian theory, organisms with all their complex systems came into existence via natural selection acting on randomly arising, small-scale variations and mutations. As Frazzetta understood, this evolutionary mechanism necessarily transforms organisms gradually, with modifications parceled into increments “as a sort of continuous change, where one structural condition melts gradually into another.”1 Frazzetta had his doubts, however. As an expert in functional biomechanics—studying how animals actually work—he had dissected the skulls of rare snakes found only on the island of Mauritius, in the Indian Ocean. These snakes, called bolyerines, are boa-like but have an anatomical specialization found in no other vertebrate. Their maxilla, the tooth-bearing bone of the upper jaw, is divided into two segments, linked by a flexible joint and serviced by many specialized nerves, extra bones, tissues, and differently arranged ligaments. This unique trait allows the snakes to bend the front half of their upper jaw backwards when they attack prey (see Fig. 12.1). Could this complex system of bones, joints, tissues, and ligaments have evolved gradually? “A movable joint dividing the maxilla into two segments,” observed Frazzetta, “seems to have either a presence or absence, with no intermediate to connect the two conditions.”2 That is, either the maxilla occurs as one bone (as it does in every other vertebrate) or as two segments with all the accompanying joints, bon...

3 Or as Frazzetta himself observed, “I thus find it difficult to envision a smooth transition from a single maxilla to the divided condition seen in bolyerines.”4 Yet because the intermediate forms would not be viable, building a bolyerine jaw would require all the necessary parts—the jointed maxilla, the adjoining ligaments, and the necessary muscles and tissues—arising together. Yet the problem for neo-Darwinian theory, Frazzetta realized, extended well beyond the anatomical peculiarities of rare snakes. As a young evolutionary biology professor, he had studied complex features in a wide variety of species. He knew that almost any biological structure of interest—the inner ear, the amniotic egg, eyes, olfactory organs, gills, lungs, feathers, the reproductive, circulatory, and respiratory systems—possesses multiple necessary components. To change such systems requires altering each of the many independent parts upon which their functions are based. This cannot be done willy-nilly. For example, changing any of the three bones of the mammalian inner ear—the incus, stapes, or malleus—will perforce require corresponding changes in the other bones and in other parts of the ear as well, such as the tympanic membrane or the cochlea. Complex biological systems depend for their functions on tens or hundreds of such independent, yet jointly necessary parts. As the number of necessary components increases, the requisite number of coordinated changes increases too, rapidly driving u...

And that was the problem, as Frazzetta understood it. Any system that depends for its function on the coordinated action of many parts could not be changed gradually without losing function. But in the neo-Darwinian scheme of things, natural selection acts to preserve only functional advantages. Changes that result in death or reduced function will not be preserved. The integrated complexity of many biological systems thus imposes limitations on the evolutionary process—limitations that human engineers do not face when they design complex integrated systems.

When modifying the design of a machine, an engineer is not bound by the need to maintain a real continuity between the first machine and the modification. . . . But in evolution, transitions from one type to the next presumably involve a greater continuity by means of a vast number of intermediate types. Not only must the end product—the final machine—be feasible, but so must be all the intermediates. The evolutionary problem is, in a real sense, the gradual improvement of a machine while it is running!

Darwin famously employed this strategy to explain the origin of the eye, asking his readers to imagine a series of incremental, advantageous changes to a simple “nerve sensitive to light.”

Frazzetta’s concerns about the adequacy of the neo-Darwinian mechanism, like Eden’s, turned on the growing appreciation of the nature and importance of genetic information. Though biologists then (as now) didn’t fully understand how genetic information in DNA correlates or “maps” to these higher-level complex morphological structures, by 1975 they did know that many hundreds of genes can be involved in coding for a single complex integrated structure. Thus, altering the anatomical structure of the mammalian ear or the vertebrate eye, for example, would involve altering the genes that code for its constituents, which implies, most implausibly, that multiple coordinated mutations would occur virtually simultaneously.

“Phenotypic alteration of integrated systems requires an improbable coincidence of genetic (and hence, heritable phenotypic) modifications of a tightly specified kind.”8 Yet the extreme specificity of the fit of the components and the functional dependence of the whole system on this fit imply limits to allowable genetic change. Genetic change affecting any one of the necessary components, unless matched by many corresponding—and vastly improbable—genetic changes, will result in functional loss and often death.

“We are still left with the unabating need to explain evolutionary changes in systems that have the operational integration characteristic of things we recognize as ‘machines.’ ”9 At the time, the doubts he expressed gained little traction in the evolutionary biology community, because neo-Darwinian evolutionary biologists assumed that mutation and selection had nearly unlimited...

As long as mutations generate a continuous supply of new traits, any system, however complex, can be built one trait at a time—trait upon trait—via the creative power of natural selection. Or so the story goes. Confidence in these mathematical models (and their underlying assumptions) led many neo-Darwinists to disregard the need to give detailed accounts of the specific evolutionary pathways by which complex systems might have arisen.

Paul Ehrlich

The neo-Darwinian focus on mathematical modeling helps to explain why mainstream evolutionary biologists haven’t worried about the problem of the origin of new genes and proteins or the problem of combinatorial inflation, discussed in Chapters 9 and 10. Many contemporary evolutionary biologists, like the founders of population genetics, assumed that some mechanism for building new genes already existed. Indeed, they assumed that new traits (and the genes for building them) can arise as the result of even single mutations (or a series of such mutations that each confer a small, incremental, selectable advantage). Thus, the mathematical expression of neo-Darwinian theory seemed to certify the plausibility of even large-scale evolutionary changes—again, provided these changes could occur one mutation at a time. But what if there are systems in living organisms that cannot be built one mutation at a time, and instead must be built by simultaneous coordinated changes?

Many evolutionary biologists have simply regarded mathematical challenges to the creative power of the mechanism, coming as they mostly do from scientists and engineers in other fields, as exotic or irrelevant. That has begun to change. And it has begun to change in a way that has not only introduced a new mathematical challenge to the creative power of the neo-Darwinian mechanism, but also in a way that indirectly confirms Axe’s key insight about the rarity of genes and proteins.

In the last decade, developments in molecular genetics and population genetics have exposed a connection between the problem of the origin of new genes and proteins and the origin of complex adaptations, a connection first perceived by Tom Frazzetta back in 1975. As more biologists have recognized that connection, they too have begun to share Frazzetta’s doubt.

The neo-Darwinian synthesis was formulated during the 1930s before the elucidation of the structure of DNA. Biologists at that time did not yet understand the nature, structure, or precise location of genetic information.11 They did not associate genes with long strings of nucleotide bases along the spine of the DNA molecule. They did not think of genes as long sections of digital code stored in complex biomacromolecules. Instead, after Mendel, but before Watson and Crick, genes were defined operationally as those entities, associated with chromosomes, that produced specific visible or selectable anatomical traits, such as eye color or beak shape. The architects of neo-Darwinism working in the 1930s reformulated evolutionary theory to emphasize the importance of mutations as the source of genetic variation. It followed, therefore, that the mutations—which they regarded as the source of heritable variation—must operate on genes. Not knowing the nature of genes, they also assumed that a single mutation could alter a gene in such a way as to produce a new trait. The equations of population genetics are predicated upon this assumption. The rate of mutation thus emerges as an important factor in computing the amount of evolutionary change that can occur in any given population. If every individual mutation can produce a new, potentially selectable trait, then the rate at which such variation accumulates partially determines how much change can occur in a given time. After 1953,...

After the 1950s, evolutionary biologists no longer assumed that single mutations would necessarily generate whole new traits. That left a critical assumption of population genetics essentially undefended. For many evolutionary biologists, the theory of gene duplication closed that conceptual gap. After the theory was formulated, many evolutionary biologists thought that a mechanism had been discovered by which sections of genetic text could accumulate multiple changes without compromising the fitness of an organism, thus ensuring the eventual production of new genes and a steady supply of new traits. So when Frazzetta confronted evolutionary biology with the problem of complex adaptations in the mid-1970s, most neo-Darwinian biologists responded with a collective yawn, if they noticed it at all. His challenge was no challenge after all. So implied the math of population genetics—provided its assumptions about the ease with which new mutations could generate new traits were valid. But were they valid? Could a series of separate mutations generate the new genes necessary to build new proteins and new traits, or did building genes require multiple coordinated mutations?

In his chapter in the 1909 anthology Darwin and Modern Science, the British geneticist William Bateson wryly described how this widespread assumption prevented evolutionary biologists from confronting the real difficulty of explaining the origin of complex adaptations: By suggesting that the steps through which an adaptive mechanism arises are indefinite and insensible, all further trouble is spared. While it could be said that species arise by an insensible and imperceptible process of variation, there was clearly no use in tiring ourselves by trying to perceive that process. This labour-saving counsel found great favor.12

One of the first prominent evolutionary biologists to consider the possibility that building new genes and proteins might require multiple coordinated mutations was John Maynard Smith. Maynard Smith worked as an aeronautical engineer during World War II, but then took up the formal study of evolutionary biology after the war. He eventually helped to found the University of Sussex, where he also served as a distinguished professor of biology until the mid-1980s.13 In 1970, Maynard Smith wrote an article in Nature responding to an earlier article by Frank Salisbury, a biologist from Utah State University. Salisbury had raised questions about whether random mutations could explain the specificity of the arrangement of nucleotide bases necessary to produce functional proteins. Salisbury worried, following discussions at Wistar, that the probability of random mutations generating functional arrangements of bases or amino acids was prohibitively low. According to Salisbury’s calculations, “The mutational mechanism as presently imagined could fall short by hundreds of orders of magnitude of producing, in a mere four billion years, even a single required gene.”14

While admitting that the origin of the first proteins remained a mystery, he suggested that one protein could evolve into another as the result of small incremental changes in amino-acid sequences, provided each sequence maintained some function at each step along the way. Maynard Smith compared protein-to-protein evolution to changing one letter in an English word in order to generate a different word (while at each step generating a different meaningful word). He used this example to convey how he thought protein evolution might work:

WORD → WORE → GORE → GONE → GENE He explained: The words [in this analogy] represent proteins; the letters represent amino acids; the alteration of a single letter corresponds to the simplest evolutionary step, the substitution of one amino acid for another; and the requirement of meaning corresponds to the requirement that each unit step in evolution should be from one functional protein to another.15 As a self-professed “convinced Darwinist,” Maynard Smith realized that natural selection and random mutation could only build new biological structures from preexisting structures if each intermediate structure along the way conferred some adaptive advantage. He thought that this requirement applied as much to the evolution of new genes and proteins as it did to the evolution of new phenotypic traits or larger-scale anatomical structures.16 Nevertheless, the essentially digital or alphabetic character of the genetic information that directs protein synthesis suggested a problem to Maynard Smith. How, he asked, could one gene or protein evolve into another if such a transformation required multiple simultaneous changes in the bases of the genetic text (or arrangement of amino acids)? If building new genes required multiple coordinated mutations, then the probability of generating a new gene or protein would drop precipitously, since such a transformation would require not just one improbable mutational event, but two or three or more, occurring more or less at once. Here’s ho...

In the end, however, he concluded that such mutations were so improbable that they must not have played a significant role in the evolution of novel structures. As he explained, “Such double steps . . . may occasionally occur, but are probably too rare to be important in evolution.”18

Not until the first decade of the twenty-first century would biologists confront the challenge of making a rigorous quantitative analysis of the plausibility of protein-to-protein evolution.

Behe had established himself as a prominent critic of neo-Darwinism by arguing that the neo-Darwinian mechanism did not provide an adequate explanation for the origin of functionally integrated “irreducibly complex” molecular machines.

They considered the plausibility of the main neo-Darwinian model of gene evolution in which evolutionary biologists envision new genes arising by gene duplication and subsequent mutations in the duplicated gene.

Whereas Maynard Smith saw the need for multiple coordinated mutations as a potential problem, one that ultimately needn’t trouble evolutionary biologists, Behe and Snoke argued that evolutionary biologists do need to worry about it, and they quantified its severity. Behe and Snoke first noted that many proteins, as a condition of their function, require unique combinations of amino acids interacting in a coordinated way. For example, ligand binding sites on proteins—places where small molecules bind to large proteins to form larger functional complexes—typically require a combination of several amino acids. Behe and Snoke argued that in such cases the combinations of amino acids would have to arise in a coordinated fashion since the capacity for ligand binding depends on all the necessary amino acids being present together. In support of this inference, they cited an authoritative textbook, Molecular Evolution, by University of Chicago evolutionary biologist Wen-Hsiung Li. In it, Li notes that evolving ligand binding capacity in proteins such as hemoglobin may require “many mutational steps,”21 even though the first steps on the way to building such capacity would confer no selective advantage. As Li explains, “Acquiring a new function may require many mutational steps, and a point that needs emphasis is that the early steps might have been selectively neutral [non-advantageous] because the new function might not be manifested until a certain number of steps had already oc...

Behe and Snoke point out that this observation implies that a series of separate mutations could not generate a ligand binding function in a protein that previously did not have this capacity, since individual amino-acid changes would initially confer no selectable advantage on the protein lacking this function. Instead, evolving ligand binding capability would require multiple coordinated mutations. Behe and Snoke make a similar argument about the requirements for the evolution of protein-to-protein interactions. They note that for proteins to interact with each other in specific ways, typically at least several individually necessary amino acids must be present in combination in each protein, again, suggesting the need for multiple coordinated mutations.

Behe and Snoke used the principles of population genetics to assess the likelihood of various numbers of coordinated mutational changes occurring in a given period of time. They asked: Is it probable that there was enough time in evolutionary history to generate coordinated mutations? If so, how many coordinated mutations is it reasonable to expect in a period of time given various population sizes, mutation rates, and generation times? Then, for different combinations of these various factors, they assessed how long it would typically take to generate two or three or more coordinated mutations. They determined that generally the probability of multiple mutations arising in close (functionally relevant) coordination to each other was “prohibitively” low—it would likely take an immensely long time, typically far longer than the age of the earth.

The Powerball website lists the probability of matching all six balls at roughly 1 in 175 million.

“If the odds of winning are one in a hundred million, and if a million people play every time, then it will take on average about a hundred drawings for someone to win.” If there are about a hundred drawings per year, with a million people playing per drawing, “then it would take about a year before someone won. But if there were only one drawing per year, on average it would take a century to hit the jackpot.”23 More frequent drawings produce shorter waiting times. Less frequent drawings tend to require longer waiting times. Similarly, more players will decrease the average time necessary to produce a winner, while fewer players result in longer waits. Similar mathematical principles apply when calculating the expected waiting times for the evolution of biological features by mutation and selection. Biologists first need to assess the complexity of the system—or its inverse, the improbability of the feature occurring. As in Powerball, however, knowing the probability of an event by itself does not allow someone to calculate how long it will likely take for that event to occur. Such a calculation requires also knowing the size of the population (equivalent to how many people are playing Powerball) and how frequently new genetic sequences arise (equivalent to how frequently drawings are held). In Powerball, a new sequence of numbers arises in every drawing. But when organisms reproduce, they do not always generate a new sequence of nucleotide bases in their individual genes...

The Powerball website lists the probability of winning a $4 prize at just 1 in 55, the probability of winning $1 million dollars at roughly 1 in 5 million, and the probability of winning the jackpot as 1 in 175 million (see Fig. 12.3).25