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

by Stephen C. Meyer

Neo-Darwinists have long assumed that biological evolution works something like matching one number in Powerball. In their view, natural selection acts to reward or preserve small but relatively probable changes in gene sequences—like winning the small but more likely $4 prize in Powerball over and over again. They assume the mutation and selection mechanism doesn’t depend on winning extremely unlikely “prizes” (like the whole Powerball jackpot) all at once. But what if, to produce a functional advantage at the genetic level, the mutation and selection mechanism had to generate the biological equivalent of all six (or more) correct balls in the Powerball lottery with no reward for guessing a smaller number of balls correctly first? Clearly, the probability of this would be extremely small. And the waiting time for winning such a lottery could become prohibitively long.

That brings us back to Behe and Snoke’s conclusion. In their 2004 paper, they argued that generating a single new protein will often require many improbable mutations occurring at once. They took into account the improbability of multiple functionally necessary mutations appearing together—the equivalent of needing to get a Powerball ticket matching several numbers to win any money at all. Then they sought to determine how long it would take and/or how large the population sizes would need to be to generate a new gene via multiple coordinated mutational changes—the genetic equivalent of the “jackpot scenario.”

Behe and Snoke found that if generating a new gene required multiple coordinated mutations, then the waiting time would grow exponentially with each additional necessary mutational change. They also assessed how population sizes affected how long it would take to generate new genes, if multiple coordinated mutations were necessary to produce those genes. They found, not surprisingly, that just as larger populations diminished expected waiting times, smaller populations dramatically increased them. More important, they found that even if building a new gene required just two coordinated mutations, the neo-Darwinian mechanism would likely either require huge population sizes or extremely long waiting times or both. If coordinated mutations were necessary, then evolution at the genetic level faced a catch-22: for the standard neo-Darwinian mechanism to generate just two coordinated mutations, it typically either needed unreasonably long waiting times, times that exceeded the duration of life on earth, or it needed unreasonably large population sizes, populations exceeding the number of multicellular organisms that have ever lived. To get population sizes that were reasonable, they had to have waiting times that were unreasonable. To get waiting times that were reasonable, they had to have population sizes that were unreasonable. As they put it, either way the “numbers seem prohibitive.”26 Behe and Snoke found that mutation and selection could generate two coordinated mutation...

Behe and Snoke found that if generating a new functional gene or trait required more than two coordinated mutations, then excessively long waiting times were necessary regardless of the size of the population. If three or more coordinated mutations were necessary, their calculations generated no “sweet spots” at all. Thus, they concluded that “the mechanism of gene duplication and point mutation alone would be ineffective, at least for multicellular . . . species.”28

Two recent scientific publications tell this story. First, in 2007, Michael Behe published a book, The Edge of Evolution, amplifying the results of his 2004 paper with David Snoke. Using public-health data about a genetic trait—resistance to the antimalarial drug chloroquine in the one-celled organism that causes malaria—Behe provided another line of evidence and argument to support the conclusion that multiple coordinated mutations are often necessary to produce even minor genetic adaptations. Based on public-health data, Behe determined that resistance to chloroquine only arises once in every 1020 malaria-causing cells. Behe inferred, by working the problem backwards, that the trait probably required multiple—though not necessarily coordinated—mutations to develop. He called this trait a “chloroquine complexity cluster,” or a “CCC.”29 Behe wanted to explore what he called the “edge of evolution,” the limits to the creative power of mutation and selection at the genetic level. Having established that this trait could arise by random mutation in a reasonably short period of time, he wondered how much time would be required to produce traits of greater complexity in populations of various sizes. He asked his readers to consider a hypothetical genetic trait twice as complex as a CCC cluster—a feature requiring the origin of two coordinated traits, each as complex as a CCC. In other words, Behe wondered how long it would take to develop a hypothetical trait that required two ...

Behe showed that the problem of coordinated mutations was particularly acute for longer-lived organisms with small population sizes—organisms such as mammals or, more specifically, human beings and their presumed prehuman ancestors. Behe estimated, based upon relevant mutation rates, known human population sizes, and generation times, the time required for two coordinated mutations to occur in the hominid line. He calculated that producing even such a modest evolutionary change would require many hundreds of millions of years. Yet, humans and chimps are thought to have diverged from a common ancestor only 6 million years ago. Behe’s calculation implied that the neo-Darwinian mechanism does not have the capacity to generate even two coordinated mutations in the time available for human evolution—and thus does not explain how humans arose. Here the story gets really interesting. Soon after the publication of The Edge of Evolution, two Cornell University mathematical biologists, Rick Durrett and Deena Schmidt, both defenders of neo-Darwinism, attempted to refute Behe’s conclusion by making their own calculations. Their paper, “Waiting for Two Mutations: With Applications to Regulatory Sequence Evolution and the Limits of Darwinian Evolution,” also applied a model based upon population genetics to calculate the amount of time necessary to generate two coordinated mutations in the hominid line. Although they calculated a shorter waiting time then Behe did, their result neverthe...

Behe specifically showed that several molecular machines within cells (such as the cilium and intraflagellar transport system, and the bacterial flagellar motor33) require the coordinated interaction of multiple protein parts in order to maintain their function.

Some neo-Darwinists have proposed a model of protein evolution known as “co-option.” In this model, a protein that performs one function is transformed, or “co-opted” to perform some other function.

the difficulty scientists have had in showing any real change of protein function to be feasible,

that Kbl2 and BioF2 were about as close in sequence and structure as any two known proteins that performed different functions.

Their results were unambiguous. They found that they could not induce, with either one or a small number of amino acids, the change in function they sought. In fact, they found that they could not get Kbl2 to perform the function of BioF2, even if they mutated larger numbers of amino acids in concert—that is, even if they made many more coordinated mutations than could plausibly occur by chance in all of evolutionary history. Although their attempts to convert Kbl2 to perform the function of BioF2 failed, their experiment did not. It allowed them to establish experimentally for the first time that the co-option hypothesis of protein evolution lacks credibility—simply too many coordinated mutations would be required to convert one protein function to another, even in the case of extremely similar proteins. That implied that generating new genes and proteins would require multiple coordinated mutations, and thus, the waiting times that Behe and Snoke had calculated do present a problem for neo-Darwinian theory.

Axe developed a refined population-genetics mathematical model to calculate various waiting times. His results roughly confirmed the previous calculations of Behe and Snoke. He found, for example, that if he took into account the probable fitness cost to an organism of carrying unnecessary gene duplicates (as was necessary to give the evolution of a new gene a reasonable chance), that the probable waiting time for even three coordinated mutations exceeded the duration of life on earth. He therefore effectively determined an upper bound of two for the number of coordinated mutations that could be expected to occur in a duplicate gene during the history of life on earth (taking into account the negative effects of carrying gene duplicates in the evolutionary process). He also calculated six coordinated mutations as an upper bound, neglecting the fitness cost of carrying gene duplicates. Nevertheless, in their experiments, he and Gauger could not induce a functional change in a single gene with more than six coordinated mutations. So, even that more generous—and, again, unrealistically generous upper bound—does little to render the co-option hypothesis credible. Indeed, Axe and Gauger’s experiments showed that the smallest realistically conceivable step exceeded what was plausible given the time available to the evolutionary process. In their words, “evolutionary innovations requiring that many changes . . . would be extraordinarily rare, becoming probable only on timescales ...

By showing the implausibility of the co-option model of protein evolution and the need for multiple coordinated mutations in order to generate multisite features in proteins, Axe and Gauger confirmed that genes and proteins themselves represent complex adaptations—entities that depend upon the coordinated interaction of multiple subunits that must arise as a group to confer any functional advantage. The need for coordinated mutations means that evolutionary biologists cannot just assume that mutations will readily generate new genes and traits, as neo-Darwinists have long presupposed.

generating the number of multiple coordinated mutations needed to produce even one new gene or protein is unlikely to occur within a realistic waiting time. Thus, these biologists establish the implausibility of the neo-Darwinian mechanism as a means of generating new genetic information.

If functional sequences are rare in sequence space, it stands to reason that finding them by purely random and undirected means will take a long time. Moreover, waiting times increase exponentially with each additional necessary mutation. Thus, long waiting times for the production of new functional genes and proteins is exactly what we should expect if indeed functional genes and proteins are rare, and if coordinated mutations are necessary to produce them.

the neo-Darwinian mechanism cannot generate the information necessary to build new genes, let alone a new form of animal life, in the time available to the evolutionary process.

Rarely have the implications of a Nobel Prize–winning scientific discovery received so little notice.

Starting in the autumn of 1979, at the European Molecular Biology Laboratory in Heidelberg, two venturesome young geneticists, Christiane Nüsslein-Volhard and Eric Wieschaus (see Fig. 13.1), generated thousands of mutations to investigate the genomes of tens of thousands of fruit flies (species: Drosophila melanogaster). They hoped to get them to divulge the secrets of embryological development. In technical jargon, Nüsslein-Volhard and Wieschaus performed “saturation mutagenesis” experiments. After feeding male flies the potent mutation-causing chemical (i.e., mutagen) ethyl methane sulphonate (EMS), Nüsslein-Volhard and Wieschaus bred the males with virgin females. They then examined the offspring larvae for visible defects. In generating many thousands of mutants, thereby “saturating” the Drosophila genome, Nüsslein-Volhard and Wieschaus induced variations in the small subset of genes that specifically regulate embryonic development. These regulatory genes normally control the expression of many other genes that build the fly embryo, progressively subdividing it into regions that will become the head, thorax, and abdomen of the adult fly. The EMS mutagen disrupts DNA replication, thereby mutating genes. These mutations affect the process of development, leaving visible defects in the fly larvae.

“This work was revolutionary,” University of Cambridge geneticist Daniel St. Johnston explained, “because it was the first mutagenesis in any multicellular organism that attempted to find most or all of the mutations that affect . . . the essential patterning genes that are used throughout development.”2

But the mutant fruit flies obtained by Nüsslein-Volhard and Wieschaus tell another story—one less widely known, but one containing important clues for the unsolved mystery of the origin of animal body plans.

Without exception, the mutants he studied perished as deformed larvae long before achieving reproductive age. “No, dead is dead,” he joked, “and you can’t be more dead.”3

“The problem is, we think we’ve hit all the genes required to specify the body plan of Drosophila,” he said, “and yet these results are obviously not promising as raw materials for macroevolution. The next question then, I guess, is what are—or what would be—the right mutations for major evolutionary change? And we don’t know the answer to that.”4

If mutating the genes that regulate body-plan construction destroy animal forms as they develop from an embryonic state, then how do mutations and selection build animal body plans in the first place?

But even if mutation and selection could generate fundamentally new genes and proteins, a more formidable problem remains. To build a new animal and establish its body plan, proteins need to be organized into higher-level structures. In other words, once new proteins arise, something must arrange them to play their parts in distinctive cell types. These distinctive cell types must, in turn, be organized to form distinctive tissues, organs, and body plans. This process of organization occurs during embryological development. Thus, to explain how animals are actually built from smaller protein components, scientists must understand the process of embryological development.

As much as any other subdiscipline of biology, developmental biology has raised disquieting questions for neo-Darwinism. Developmental biology describes the processes, called ontogeny, by which embryos develop into mature organisms. Within the past three decades the field has dramatically advanced our understanding of how body plans arise during ontogeny.

Much of this new knowledge has come from studying so-called model systems—organisms that biologists can easily mutate in the lab, such as the fruit fly Drosophila and the nematode Caenorhabditis elegans. Although the exact details of animal development can vary in bewildering ways depending on the species, all animal development exemplifies a common imperative: start with one cell, end with many different cells. In most animal species, development begins with the fertilized egg. Once the egg divides into its daughter cells, becoming an embryo, the organism begins heading toward a well-defined target, namely, an adult form that can reproduce. Arriving at that distant target requires the embryo to produce many specialized cell types, in the correct positions and at the right time.

For example, the specialized digestive proteins that service the cells lining the adult intestine differ from proteins expressed in a neuron found in the nerve tract of a limb. They must differ because each performs dramatically different functions. So, during development, the appropriate genes must be turned on, or “up-regulated,” and turned off, or “down-regulated,” to ensure the production of the correct protein products at the right time and in the right cell types. Specific proteins play active roles in regulating the expression of genes for building other proteins. The protein actors playing these coordinating roles are known as transcriptional regulators (TRs) or transcription factors (TFs). TRs (or TFs) usually bind directly to specific sites in DNA, either preventing (repressing) or enabling (activating) the transcription of specific genes into RNA. TRs or TFs convey instructions about which genes to turn on or turn off. Their three-dimensional geometries exhibit characteristic DNA-binding features, including a specific domain of 61 amino acids that wraps around the DNA double helix. Other transcription factors include the zinc finger and leucine zipper motifs that also bind to DNA. Transcriptional regulators and factors are themselves controlled by complex circuits and signals transmitted by other genes and proteins, the overall complexity and precision of which is breathtaking. Painstaking genetic research—performed by Nüsslein-Volhard and Wieschaus and many oth...

To create significant changes in the forms of animals requires attention to timing. Mutations in genes expressed late in the development of an animal will affect relatively few cells and architectural features. That’s because by late in development the basic outlines of the body plan have already been established.6 Late-acting mutations therefore cannot cause any significant or heritable changes in the form or body plan of the whole animal.

Thus, mutations that are expressed early in the development of animals have probably the only realistic chance of producing large-scale macroevolutionary change.

Yet from the first experiments by geneticist T. H. Morgan systematically mutating fruit flies early in the twentieth century until today, as many model species have been subjected to mutagenesis, developmental biology has shown that mutations affecting body-plan formation expressed early in development inevitably damage the organism.

11 (See Fig. 13.2, for examples.) As one of the founders of neo-Darwinism geneticist R. A. Fisher noted, such mutations are “either definitely pathological (most often lethal) in their effects,” or they result in an organism that cannot survive “in the wild state.”12

Normal development in any animal can be represented as an expanding network of decisions, where the earliest (upstream) decisions have greater impact than those occurring later. Regulatory genes and their DNA-binding protein products help to control this unfolding network of decisions—such that if regulatory proteins are altered or destroyed by mutation, the effects cascade downstream into the whole developmental process. The earlier the failure, the more widespread the destruction. Geneticist Bruce Wallace explains why early-acting mutations are thus overwhelmingly likely to disrupt animal development. “The extreme difficulty encountered,” he observes, “when attempting to transform one organism into another . . . still functional one lies in the difficulty in resetting a number of the many controlling switches in a manner that still allows for the individual’s orderly (somatic) development.”13 Nüsslein-Volhard and Wieschaus discovered this problem in experiments performed on fruit flies after their first Nobel Prize–winning efforts. In these later experiments they studied protein molecules that influence the organization of different types of cells early in the process of embry...

This problem has led to what Georgia Tech geneticist John F. McDonald has called a “great Darwinian paradox.”16 He notes that the genes that are obviously variable within natural populations seem to affect only minor aspects of form and function—while those genes that govern major changes, the very stuff of macroevolution, apparently do not vary or vary only to the detriment of the organism. As he puts it, “Those [genetic] loci that are obviously variable within natural populations do not seem to lie at the basis of many major adaptive changes, while those loci that seemingly do constitute the foundation of many if not most major adaptive changes are not variable within natural populations.”

17 In other words, the kind of mutations the evolutionary process would need to produce new animal body plans—namely, beneficial regulatory changes expressed early in development—don’t occur. Whereas, the kind that it doesn’t need—viable genetic mutations in DNA generally expressed late in development—do occur. Or put more succinctly, the kind of mutations we need for major evolutionary change we don’t get; the kind we get we don’t need. My Discovery Institute colleague Paul Nelson (see Fig. 13.3), a philosopher of biology who specializes in evolutionary theory and developmental biology, summarizes the challenge to neo-Darwinism posed by animal development as three premises: 1. Animal body plans are built in each generation by a stepwise process, from the fertilized egg to the many cells of the adult. The earliest stages in this process determine what follows. 2. Thus, to evolve any body plan, mutations expressed early in development must occur,

must be viable, and must be stably transmitted to offspring. 3. Such early-acting mutations of global effect on animal development, however, are those least likely to be tolerated by the embryo and, in fact, never have been tolera...

the scientific literature offers no examples of viable mutations affecting early animal development and body-plan formation (Premise 3, on previous page) and also that the macroevolution of novel animal form requires just such early-acting mutations (Premise 2, on previous page).

“If the only kind of mutations that can conceivably produce enough morphological change to alter whole body plans never causes beneficial and heritable changes, then it is difficult to see how mutation and selection could ever produce new body plans in the first place.”

Thus, he concludes: Research on animal development and macroevolution over the last thirty years—research done from within the neo-Darwinian framework—has shown that the neo-Darwinian explanation for the origin of new body plans is overwhelmingly likely to be false—and for reasons that Darwin himself would have understood.

Indeed, Darwin himself insisted that “nothing can be effected” by natural selection, “unless favorable variations occur.”20 Or as Danish evolutionary biologist Søren Løvtrup succinctly explains: “Without variation, no selection; without selection, no evolution. This assertion is based on logic of the simplest kind. . . . Selection pressure as an evolutionary agent becomes void of sense unless the availability of the proper mutations is assumed.”21 Yet the “proper” kind of mutations—the mutations that produce favorable changes to early-acting, body-plan–shaping, regulatory genes—do not occur. Microevolutionary change is insufficient; macromutations—large-scale changes—are harmful. This paradox has beset Darwinism from its inception, but discoveries about the genetic regulation of development in animals have made this paradox more acute and cast serious doubt on the efficacy of the modern neo-Darwinian mechanism as an explanation for the new body plans that arise in the Cambrian period.

Another line of research in developmental biology has revealed a related challenge to the creative power of the neo-Darwinian mechanism. Developmental biologists have discovered that many gene products (proteins and RNAs) needed for the development of specific animal body plans transmit signals that influence the way individual cells develop and differentiate themselves. Additionally, these signals affect how cells are organized and interact with each other during embryological development. These signaling molecules influence each other to form circuits or networks of coordinated interaction, much like integrated circuits on a circuitboard. For example, exactly when a signaling molecule gets transmitted often depends upon when a signal from another molecule is received, which in turn affects the transmission of still others—all of which are coordinated and integrated to perform specific time-critical functions. The coordination and integration of these signaling molecules in cells ensures the proper differentiation and organization of distinct cell types during the development of an animal body plan. Consequently, just as mutating an individual regulatory gene early in the development of an animal will inevitably shut down development, so too will mutations or alterations in the whole network of interacting signaling molecules destroy a developing embryo.

During the life cycle of an organism, the genomes of these specialized cells express only a small fraction of their DNA at any given time and produce different RNAs as a result. These facts strongly suggest that some animal-wide system of genetic control functions to turn specific genes on and off as needed throughout the life of the organism—and that such a system functions during the development of an animal from egg to adult as different cell types are being constructed. When they proposed their theory in 1969, Britten and Davidson acknowledged that “little is known . . . of the molecular mechanisms by which gene expression is controlled in differentiated cells.”24 Nevertheless, they deduced that such a system must be at work. Given: (1) that tens or hundreds of specialized cell types arise during the development of animals, and (2) that each cell contains the same genome, they reasoned (3) that some control system must determine which genes are expressed in different cells at different times to ensure the differentiation of different cell types from each other—

some system-wide regulatory logic must oversee and coordinate the expression of the genome.

Davidson has shown that the nonprotein-coding regions of DNA that regulate and control gene expression and the protein-coding regions of the genome together function as circuits. These circuits, which Davidson calls “developmental gene regulatory networks” (or dGRNs) control the embryological development of animals.

Figure 13.4a shows the urchin embryo as it appears six hours after development has begun (top left of diagram). This is the 16-cell stage, meaning that four rounds of cell division have already occurred (1 → 2 → 4 → 8 → 16). As development proceeds in the next four stages, both the number of cells and the degree of cellular specialization increases, until, at 55 hours, elements of the urchin skeleton come into focus. Figure

This process does not happen fortuitously in the sea urchin but via highly regulated and precise control systems, as it does in all animals. Indeed, even one of the simplest animals, the worm C. elegans, possessing just over 1,000 cells as an adult, is constructed during development by dGRNs of remarkable precision and complexity. In all animals, the various dGRNs direct what Davidson describes as the embryo’s “progressive increase in complexity”—an increase, he writes, that can be measured in “informational terms.”28 Davidson notes that, once established, the complexity of the dGRNs as integrated circuits makes them stubbornly resistant to mutational change—a point he has stressed in nearly every publication on the topic over the past fifteen years. “In the sea urchin embryo,” he points out, “disarming any one of these subcircuits produces some abnormality in expression.”29 Developmental gene regulatory networks resist mutational change because they are organized hierarchically. This means that some developmental gene regulatory networks control other gene regulatory networks, while some influence only the individual genes and proteins under their control. At the center of this regulatory hierarchy are the regulatory networks that specify the axis and global form of the animal body plan during development. These dGRNs cannot vary without causing catastrophic effects to the organism. Indeed, there are no examples of these deeply entrenched, functionally critical circuits v...

Davidson’s work highlights a profound contradiction between the neo-Darwinian account of how new animal body plans are built and one of the most basic principles of engineering—the principle of constraints. Engineers have long understood that the more functionally integrated a system is, the more difficult it is to change any part of it without damaging or destroying the system as a whole.

The system of gene regulation that controls animal-body-plan development is exquisitely integrated, so that significant alterations in these gene regulatory networks inevitably damage or destroy the developing animal.32 But given this, how could a new animal body plan, and the new dGRNs necessary to produce it, ever evolve gradually via mutation and selection from a preexisting body plan and set of dGRNs? Davidson makes clear that no one really knows: “contrary to classical evolution theory, the processes that drive the small changes observed as species diverge cannot be taken as models for the evolution of the body plans of animals.”33 He elaborates: Neo-Darwinian evolution . . . assumes that all process works the same way, so that evolution of enzymes or flower colors can be used as current proxies for study of evolution of the body plan. It erroneously assumes that change in protein-coding sequence is the basic cause of change in developmental program; and it erroneously assumes that evolutionary change in body-plan morphology occurs by a continuous process. All of these assumptions are basically counterfactual. This cannot be surprising, since the neo-Darwinian synthesis from which these ideas stem was a premolecular biology concoction focused on population genetics and . . . natural history, neither of which have any direct mechanistic import for the genomic regulatory systems that drive embryonic development of the body plan.34

Darwin’s doubt about the Cambrian explosion centered on the problem of missing fossil intermediates. Not only have those forms not been found, but the Cambrian explosion itself illustrates a profound engineering problem that fossil evidence does not address—the problem of building a new form of animal life by gradually transforming one tightly integrated system of genetic components and their products into another.

DNA is not the whole story, that other sources of information are playing important roles in directing at least the early stages of animal development.

Instead, many of the scientists in their volume reported that so-called epigenetic information—information stored in cell structures, but not in DNA sequences—plays a crucial role. The Greek prefix epi means “above” or “beyond,” so epigenetics refers to a source of information that lies beyond the genes. As Müller and Newman explain in their introduction, “Detailed information at the level of the gene does not serve to explain form.”5 Instead, as Newman explains, “epigenetic” or “contextual information” plays a crucial role in the formation of animal “body assemblies” during embryological development.6