After a series of stinging defeats in the U.S. courts, one branch of biblical creationism mutated into a neocreationist movement known as intelligent design (ID) in the late 1980s. The selective pressure that led to this evolution was a single line in the 1987 Edwards decision by Supreme Court Justice William Brennan that emphasized differing views of human origins might be permissible in public schools if they were driven by “secular intent.” This led some creationists to strip the religious language and present their ideas as purely scientific. No serious person was fooled by this game of rhetorical dress-up, least of all a George W. Bush–appointed federal judge who ruled that ID is not scientific and cannot be divorced from its religious motivations (Jones 2005). (For a history of the modern ID movement, see Barbara Forrest and Paul R. Gross’s 2007 book, Creationism’s Trojan Horse: The Wedge of Intelligent Design.)
Among the most obvious signs of the pseudoscientific nature of ID is the almost complete absence of relevantly qualified experts. Its leaders are mostly theologians, philosophers, and historians, along with some physicians, engineers, and mathematicians, but there are shockingly few scientists—and even fewer biologists—and almost none have research experience in evolutionary or molecular biology. This gives them no pause in attempting to challenge the entire scientific establishment, not through original research but through books and articles published on Evolution News, the most influential ID-supporting website. Although Evolution News presents itself as a science website, it frequently covers social issues such as abortion, religious freedom, gay marriage, euthanasia, and their favorite bogeyman: Marxism.
Another sign of the pseudoscientific nature of ID proponents is their confident assertions regarding the imminent ascendance of ID. Readers are assured that scientists are confused and desperate because the end is nigh for ideas such as the big bang and the whole of evolutionary theory. Claims such as, “Given a ‘primal blueprint’ that preceded the first life, the solution can only be intelligent design,” are offered as scientific certainties even though scientists who actually study the origins of life see no such blueprint. Similarly, the past ten years have seen unprecedented discoveries that help illuminate our evolutionary past, yet, as Evolution News reports it, the discoveries of the 2010s “actually weakened the evidence for human evolution.” Not a single active paleoanthropologist would agree.
The Author
One of the very few legitimately credentialed molecular biologists who subscribes to ID is Lehigh University biochemist Michael Behe. While most ID material is unworthy of a response, Behe is a clear and engaging writer with particular command of the molecular details of life. He is neither a rube nor a neophyte and writes with a polished veneer of scientific authority. In his 1996 book Darwin’s Black Box, Behe asserts that many biological structures harbor the property of irreducible complexity, being so intricately complex with multiple interconnected parts that they could not possibly have evolved piecemeal, because the overall structure has no function until all parts are in place. To be sure, the challenge that irreducible complexity posed to the field of molecular evolution was rhetorical, not scientific, but it did draw attention to these important questions and inspired scientists to more fully flesh out the theoretical framework for how some complex molecular structures evolve and to better explain this framework to the public. For that reason, Darwin’s Black Box could be charitably seen as a contribution to science.
Every single one of Behe’s examples of irreducibly complex molecules and structures has since been thoroughly debunked by scientists. Russell Doolittle taught us how the blood clotting cascade likely evolved, with no reliance on incredibly unlikely steps (Doolittle et al. 2008). The bacterial flagellum evolved from the type III secretion machinery (Miller 2004). The vertebrate eye, and its stunningly similar counterpart in cephalopods, evolved in discrete steps of increasing, not irreducible, complexity (Lamb et al. 2008). Yet Behe continues to trumpet these examples and declares them “unchallenged.”
The Thesis
Recently, I reviewed Behe’s latest book, Darwin Devolves, for Science, the top scientific research journal in the United States (Lents et al. 2019). Behe argues that unguided random mutations serve primarily to damage genes and that doing so is occasionally good for the organism, leading to adaptation through natural selection of these damaging-but-advantageous mutations. Thus, Behe accepts that microevolution through random mutation can diversify organisms into species and genera—and perhaps families—but that something more is needed for large-scale evolutionary transitions.
Behe accepts the true age of the earth and admits that the evidence for the common ancestry of all life, including humans, is overwhelming. He sees a gap in evolutionary theory, however, in explaining the emergence of new kinds of organisms and concludes that it must require the intervention of the designer in some way. He stops short of providing evidence for the intervention and instead presents ID as the default position left standing after he shows how the unguided forces of evolution are insufficient for accounting for the origins and diversity of life.
But Behe does not show that. Darwin Devolves is mostly dedicated to explaining, often in great detail, how some high-profile examples of evolutionary research actually favor his view, rather than the interpretation of the scientists who did the work. As I will show, his discussion of every single example is misleading, sometimes egregiously so, insofar as he exaggerates the evidence that supports his view and ignores or dismisses the evidence that doesn’t. But even more damning is his near-complete omission of any discussion of the molecular and evolutionary forces that are responsible for the very phenomena he focuses on.
The Omissions
If Behe seeks to support his claim that standard evolutionary forces are insufficient to generate adaptive innovations, one would think he would dedicate quite a bit of time to discussing those forces and why scientists are wrong about them. Instead, he takes the opposite approach and either summarily dismisses them or ignores them altogether. For example, one molecular mechanism that has driven otherwise incredibly improbable evolutionary events is horizontal gene transfer, when genetic material moves from one species to another, usually through a virus (Keeling and Palmer 2008). This phenomenon is most famous for the origin of mitochondria and chloroplasts, foundational events in the evolution of complex cells, but it has been shown to occur in more mundane instances as well. For example, deer ticks gained new defenses against bacteria through genes that came from the bacteria themselves (Chou et al. 2015). While certainly not a common occurrence, horizontal gene transfer can, in one momentous instant, have profound effects in the evolutionary potential of a lineage. Despite its importance in the very concepts that Behe tackles, he does not mention this concept even once in Darwin Devolves.
Also unmentioned by Behe is exaptation, the co-opting of a structure, be it a molecule or anatomy, for a new function. For example, two of the three bones of the mammalian middle ear were co-opted from jaw bones in our reptilian ancestors. Wings, feathers, and swim bladders are other well-worn anatomical examples, but exaptation is even more prolific at the molecular scale. With the subtlest of tweaks, enzymes can catalyze different reactions, genes can be expressed in different tissues, and proteins can find new binding partners. Though Behe does not bother to address this, the molecular possibilities of exaptation are endless, particularly when gene duplication is involved. In the age of genomics, the evidence for molecular exaptation is abundant.
Speaking of gene duplication, Behe barely mentions this phenomenon either, despite it being a common first step in evolutionary innovation and diversification. When a gene has an essential function, it is tightly constrained by natural selection, and almost any mutation would cause harm and be eliminated. When a gene has been duplicated, however, the original copy can retain essential functions while the randomness of mutation and recombination fiddle around with the new one. In a powerful 2012 article, completely unmentioned by Behe, Dan Andersson and colleagues showed that gene duplication—through random mutations—can lead bacterial cells to evolve the ability to synthesize an amino acid that they were previously unable to (Näsvall et al. 2012). What Behe claims evolution can’t do, scientists have already shown it can.
Behe fails to discuss genetic recombination as well. It is beyond frustrating that Behe claims that evolution has no means to generate adaptive innovations when not only do such mechanisms exist but there is a rich literature of research on them that he simply ignores. This is part of a larger pattern by ID proponents, and Behe in particular, of ignoring the very evolutionary mechanisms that they claim do not exist. Furthermore, Darwin Devolves, like Behe’s previous two books, frequently discusses evolution and natural selection as though they are synonymous or as though natural selection is the sole engine of evolution. But natural selection is but one evolutionary force, albeit the most famous one. Mutation, genetic drift, sexual selection, recombination, horizontal gene transfer, frequency-dependent selection, and neutral theory all contribute to the toolkit of evolution. When Behe claims that natural selection alone cannot account for the rich molecular biodiversity we observe, he’s absolutely correct. Rather than looking to the designer, he should look to some of the discoveries we’ve made in the past 160 years (Darwin 1859).
The Misinterpretations
To support his view, Behe focuses his attention on several of the most high-profile examples of evolution research and attempts to reinterpret them. For example, he spends nearly an entire chapter explaining how the long-term E. coli evolution experiment (LTEE) shows how mutation and natural selection serve only to “break or blunt genes.” Richard Lenski, the architect of the LTEE for thirty years and 70,000 bacterial generations, has responded to Behe at length in a series of blog posts (Lenski 2019). It is hard to overstate just how badly Behe misinterprets the LTEE, including the fact that the controlled environment is highly artificial on purpose. There is unlimited food, static temperature, and high oxygen, and there are no competitors, pathogens, or immune system to challenge them. It is essentially an endless race at breakneck speed with strong competitive advantage for streamlining, efficiency, and growth acceleration. The LTEE was designed to detect the molecular mechanisms of evolution and has succeeded in spectacular detail, earning Lenski membership in the National Academy of Sciences (Good et al. 2017).
Behe is correct that the bacteria have adapted to this peculiar environment by ditching basically anything that slows them down, especially genes that aren’t useful in this setting. He fails to mention that plenty of gene products show enhanced function as well. Further, some of the bacteria even developed the ability to eat citrate, which is included in the growth broth to assist the absorption of iron, not as a food source (Blount et al. 2012). Behe begrudgingly admits this but dismisses it as a “sideshow” and declines to explain to his readers that the bacteria achieved this feat by reactivating a nonfunctional gene through an elaborate genomic rearrangement. Yes, this was a very improbable event, but that is exactly the point. Behe’s whole thesis is that the odds of getting new functions from unguided mutations are so low that they just don’t happen. But the LTEE proves they do.
Behe also takes aim at Darwin’s finches. He describes the remarkable work that Rosemary and Peter Grant have done on these famous birds, especially in regard to their genomes before and after a drought that briefly but significantly reduced their numbers (Grant and Grant 1993). But Behe uses the opportunity to scoff at the diversification of the finches in the first place, which took place when ancestors from the mainland found themselves stranded on the island two million years ago; their descendants have since adapted to the local habitats. He is unimpressed with their diversification into as many as eighteen species across five genera, adapting to a wide variety of foods, including mature leaves, a resource that no other bird species is known to subside on (Grant 1981).
Bizarrely, Behe compares the diversification of the finches on the Galapagos to the adaptive radiation of animals during the Cambrian explosion, more than 500 million years ago, complaining that the finches didn’t evolve “at least one crummy new phylum, class, or order.” Besides occurring over a much longer period of time and in smaller, simpler, and faster reproducing animals, the Cambrian explosion was a global diversification event that occurred when the animal kingdom was just getting started. Dry land hadn’t even been fully colonized by life yet, and new niches were opening up all over the planet. In contrast, most of the niches that Behe is surprised that the finches didn’t adapt to were already occupied when they got there. That Behe fails to grasp this reveals a poor understanding of not just ecology but evolutionary theory itself—the idea he has spent the past three decades critiquing.
The Misunderstandings
The opening section of Darwin Devolves on the evolution of polar bears may be the most erroneous example in the book. Scientists recently sequenced the full genomes of polar bears and their closest relatives, brown bears, and were able to approximate the common ancestral genome using pandas as a reference (Liu et al. 2014). Because polar bears are a relatively young species—just several hundred thousand years old—the genetic differences between them and their close cousins are relatively modest and easy to scrutinize. The scientists generated lists of polar bear genes that have undergone recent evolution. Not surprisingly, polar bears have undergone much more change than brown bears, who pretty much live in the same way that their ancestors did. As Behe sees it, the engine of polar bear evolution was the accumulation of mutations that break or diminish their genes.
The gene Behe discusses most is APOB, which codes for apolipoprotein B, the main protein responsible for clearing cholesterol from our blood following a high-fat meal. Behe declares this “the most strongly selected mutation” (it’s actually the second-most, and he later acknowledged this error). APOB has several polar bear–specific differences and has undergone strong positive selection, implying that it is important to their survival and success. This makes sense because polar bears live mostly on seal meat, a diet that is very high in saturated fat and cholesterol, much higher than their brown bear cousins who eat leaves, berries, and sometimes fish. Yet, they do not suffer much from the heart disease that we know can often result from a high cholesterol diet. Presumably, polar bear APOB must be highly optimized for handling the high cholesterol burden.
Somehow, Behe draws the opposite conclusion. This is puzzling because he notes that humans and mice with mutations that diminish the function of APOB are more prone to suffer heart disease. But polar bears do not get cholesterol-driven heart disease, so the logical inference to draw from these data is that polar bear APOB is enhanced, not diminished, by the mutations.
This is where it becomes crystal clear that Behe is blinded by his bias. Although there is no reason to conclude that polar bear APOB has been diminished based on the data alone, because that is what Behe was hoping to see, he did. In fact, he even seems to believe that the authors of the paper agree, claiming that “they determined that the mutations were very likely to be damaging.” Having read the paper from title to references, I can confidently say that the authors determined no such thing. In fact, sentences such as “adaptive changes in APOB … contribut[e] to the effective clearance of cholesterol from the blood” clearly imply that they believe that polar bear APOB is better at its job, not worse (Liu et al. 2014).
Let’s look at these mutations. First, they are not evenly distributed throughout the protein. More than half the polar bear–specific changes are in a small region called the N-terminal domain, which comprises just 22 percent of the protein. As the authors state, “This domain encodes the surface region and contains the majority of functional domains for lipid transport.” Proteins can be damaged by mutations virtually anywhere. Enhancing mutations, however, must usually be placed precisely in the domain that performs the enhanced function. Thus, the location of the APOB mutations provides more support that enhancement, not diminishment, has occurred.
The evidence on which Behe bases his claim that polar bear APOB has been damaged comes from his reading of the authors’ presentation of predictions made by a computer program called PolyPhen-2. This program can predict whether a given mutation diminishes the function of a protein based on how closely it corresponds to human genetic variation that is known to cause pathology. The PolyPhen tool was designed to identify disease-causing mutations in genomic data collected from human patients, a very different purpose than Behe’s. Trained with large data sets of human genomic variation, the program predicts whether a given mutation alters the structure of a protein and, if so, labels these possibly or probably damaging. Mutations that likely cause little or no structural change in the protein, PolyPhen calls “benign.”
On the other hand, the program has no way to predict enhancing mutations. That would require that it somehow knows all the functions of every protein and how those functions are carried out. There is no database of enhancing genetic variation in humans or any other species. Behe and others at Evolution News seem not to know this. They repeatedly insist that the program does not list any “constructive mutations” in polar bears, which is not even a possible output of the program! Importantly, what PolyPhen calls damaging (simply because the structure has been altered) could actually be a new or enhanced function, a point not lost on the study authors. In an interview with New Scientist, one of them repeated their conclusion that polar bear APOB likely makes cholesterol clearance “more efficient” (Coghlan 2014).
Behe’s reading of the polar bear research seems to begin and end with the PolyPhen predictions. This is problematic for several reasons. First, whereas damaging mutations might occur more commonly than enhancing ones, the vast majority of them don’t persist. When studying speciation, the observed mutations have already been scrutinized by natural selection. We therefore expect an enrichment of those rare mutations that PolyPhen calls damaging but are actually constructive, as seems likely to have happened here. In fact, it’s possible that none of the most strongly selected mutations are damaging. Second, other kinds of empirical data are more reliable than PolyPhen predictions, given that this is not the task the tool was designed to perform. Everything else we know about APOB points to the conclusion that the polar bear version must be well optimized.
My colleagues at Peaceful Science and I decided to look even more closely at the polar bear mutations in APOB. In a series of four discussion forums with nearly 100 contributions, eight molecular biologists pored over the data from the polar bear genome, as well as DNA sequencing data available in the NCBI database, and we scrutinized the PolyPhen-2 predictions ourselves, the kind of due diligence that Behe would have been wise to do (Swamidass 2021). We found that the picture is even murkier than it already appears. For one thing, if you substitute the human APOB protein sequence for the ancestor bear sequence as the reference in PolyPhen, only three of the nine polar bear variants are predicted to be possibly or probably damaging. (Using the human sequence as a reference is particularly useful because data from human disease-causing gene variants is how PolyPhen is trained.)
Even worse, we found that none of the nine mutations under scrutiny are actually fixed in the polar bear population. Yes, they are common, but most polar bears do not have all these individual variants. While this fact does not help us understand whether the mutations are damaging or not, it does weaken any claim that APOB was crucial in polar bear evolution. And the final nail in the coffin for this reasoning is that polar bears are resistant to atherosclerosis despite very high levels of circulating cholesterol. It may be that the polar bear adaptation to a high-fat diet has little to do with APOB whatsoever. Although we are left with an unclear picture, there is simply no hard evidence that APOB is damaged, let alone that the damage was somehow adaptive. Rather than carefully and objectively examining all available evidence, Behe simply pounced on a chart with some computer predictions that he didn’t fully understand.
The Misrepresentation
There is a very unfortunate coda to the polar bear story. After Arthur Hunt and I wrote a blog post rebutting Behe’s claims, he responded with an angry rant on Evolution News that was aimed, oddly, more at Jerry Coyne for sharing our post than at us for writing it (Behe 2019). He offered a mere two sentences of defense, posting a chart with “the relevant information” that supported “every actual undistorted claim” that he made. However, instead of presenting the actual chart from the article, he made a new chart, including only the PolyPhen predictions that support his version of polar bear evolution and removing those that argue against it.
It’s hard to overstate just how misleading this version of the chart is. In the original polar bear article, the complete list of PolyPhen predictions is in Table S7, which includes forty-eight separate mutations and the results of two PolyPhen predictions based on different genetic datasets (Liu et al. 2014). In one of them, 25 percent of the mutations are predicted to be possibly or probably damaging. Behe chose to leave those data out, preferring instead to show results from the other column, which predicted that 52 percent of the mutations might be damaging. Because this still doesn’t look like an “overwhelming tendency” of natural selection to damage genes, Behe took things further and removed all the mutations that were not predicted to be damaging—nearly half the data—to create a chart that gives the desired impression. (See accompanying figure.)
When I called out this sleight of hand on Twitter, Evolution News responded in characteristic “google swarm” fashion, publishing a whole series of articles attempting to rescue Behe’s claims about polar bear evolution while insulting me and my colleagues. They even dedicated one whole article to the “fake scandal” about the doctored chart, offering two defenses. First, they claim he was merely “saving space,” an odd concern for a web-only publication. Second, they claim that he was only providing the “relevant information,” that is, the data that shows that many of the mutations are damaging. This, too, strains credulity. When one’s entire argument is that the overwhelming trend of unguided evolution is toward the breaking or blunting of genes, one cannot deliberately obscure all contrary evidence and not expect others to cry foul. In this case, an honest and transparent presentation of the data indeed paints a very different picture from the one Behe wanted his readers to see. In the quest to be taken seriously, Evolution News is its own worst enemy.
Conclusion
Darwin Devolves is a case study in how proponents of intelligent design grasp at any evidence that they can interpret in a favorable way, while simultaneously ignoring massive amounts of opposing evidence. Although this strategy is standard fare for creationism, the age of genomics has brought nearly limitless information to our fingertips, making it easier than ever for those poorly trained in genomics to mine the data for nuggets of evidence for their claims. Of course, the wealth of data also provides the full context for debunking weak claims, as we have done with Behe’s assertions about polar bears. The careful work of analyzing and interpreting huge genomic data sets takes a great deal of time, training, and repetition. The few scientists who work within the ID framework would be well served to take the time and do the training. But let them be warned: it will be very difficult to remain within the ID framework if they do.
References
Behe, Michael J. 2019. Coyne and polar bears: Why you should never rely on incompetent reviewers. Evolution News (February 14). Available online at https://evolutionnews.org/2019/02/coyne-and-polar-bears-why-you-should-never-rely-on-incompetent-reviewers/.
Blount, Zachary D., Jeffrey E. Barrick, Carla J. Davidson, et al. 2012. Genomic analysis of a key innovation in an experimental Escherichia coli population. Nature 489(7417): 513–518.
Chou, Seemay, Matthew D. Daugherty, S. Brook Peterson, et al. 2015. Transferred interbacterial antagonism genes augment eukaryotic innate immune function. Nature 518(7537): 98–101.
Coghlan, Andy. 2014. Polar bears evolved to eat junk food. New Scientist 2969. Available online at https://www.newscientist.com/article/dn25535-zoologger-polar-bears-evolved-to-eat-junk-food/.
Darwin, Charles. 1859. On the Origin of Species by Means of Natural Selection. London: Murray.
Doolittle, Russell F., Yong Jiang, and Justin Nand. 2008. Genomic evidence for a simpler clotting scheme in jawless vertebrates. Journal of Molecular Evolution 66(2): 185–196.
Forrest, Barbara, and Paul R. Gross. 2007. Creationism’s Trojan Horse: The Wedge of Intelligent Design. Oxford University Press.
Good, Benjamin H., Michael J. McDonald, Jeffrey E. Barrick, et al. 2017. The dynamics of molecular evolution over 60,000 generations. Nature 551(7678): 45–50.
Grant, Peter R. 1981. Speciation and the adaptive radiation of Darwin’s finches: The complex diversity of Darwin’s finches may provide a key to the mystery of how intraspecific variation is transformed into interspecific variation. American Scientist 69(6): 653–663.
Grant, Barbara Rosemary, and Peter Raymond Grant. 1993. Evolution of Darwin’s finches caused by a rare climatic event. Proceedings of the Royal Society of London. Series B: Biological Sciences 251(1331): 111–117.
Jones III, John E. 2005. Tammy Kitzmiller, et al. v Dover Area School District, et al. Case No. 04cv2688-JEJ. December 20.
Keeling, Patrick J., and Jeffrey D. Palmer. 2008. Horizontal gene transfer in eukaryotic evolution. Nature Reviews Genetics 9(8): 605–618.
Lamb, Trevor D., Edward N. Pugh, and Shaun P. Collin. 2008. The origin of the vertebrate eye. Evolution: Education and Outreach 1(4): 415–426.
Lenski, Richard E. 2019. Does Behe’s ‘first rule’ really show that evolutionary biology has a big problem? Telliamed Revisted (February 15). Available online at https://telliamedrevisited.wordpress.com/2019/02/15/does-behes-first-rule-really-show-that-evolutionary-biology-has-a-big-problem/.
Lents, Nathan H., S. Joshua Swamidass, and Richard E. Lenski. 2019. The end of evolution? Science 363(6427): 590.
Liu, Shiping, Eline D. Lorenzen, Matteo Fumagalli, et al. 2014. Population genomics reveal recent speciation and rapid evolutionary adaptation in polar bears. Cell 157(4): 785–794.
Miller, Kenneth R. 2004. The flagellum unspun—the collapse of ‘irreducible complexity.’ In Philosophy of Biology: An Anthology, 439–449.
Näsvall, Joakim, Lei Sun, John R. Roth, et al. 2012. Real-time evolution of new genes by innovation, amplification, and divergence. Science 338(6105): 384–387.
Swamidass, S. Joshua. 2021. Behe meets the Peaceful Science Forum. Panda’s Thumb. Available online at https://pandasthumb.org/archives/2021/03/behe-meets-peaceful.html.