Earlier in this chapter, we saw that evolutionary mechanisms can be classified in two broad categories: mechanisms that alter the frequencies (relative abundances) of traits within populations, and mechanisms that produce new heritable traits in individual organisms. The population-influencing mechanisms—natural selection, sexual selection, genetic drift, and gene flow—are well-understood, and there is strong evidence for all four of these processes. Mathematical models and statistical techniques have been developed to study their effects.An important example is the Hardy-Weinberg equation, developed in the early 20th century by mathematician Godfrey Hardy and physician Wilhelm Weinberg. The equation describes the expected relative frequencies of alleles in a large population in “equilibrium,” that is, when the effects of selective pressures and other evolutionary mechanisms are negligible. In other words, the equation describes what to expect (in the statistical frequencies of alleles) when a population isn’t evolving. So, observed deviations from the Hardy-Weinberg equilibrium can be used to determine which genes are presently undergoing evolutionary change in a population. Some models have been quite successful, accurately predicting trends in laboratory experiments (e.g. experiments involving populations of fruit flies) and in wild animal populations.
Turning to the other category, we considered possible mechanisms by which new heritable traits might arise. The ideas involving epigenetics, evo-devo, and symbiogenesis are somewhat speculative but at least plausible. More importantly, there are at least two known mechanisms by which new traits can arise: mutation and genetic recombination. Genetic recombination obviously occurs, as demonstrated in Mendel’s experiments. Mutation also occurs frequently, as seen in the Long Term Evolution Experiment and innumerable other studies. So, there are numerous plausible mechanisms by which life could evolve, and there is abundant evidence that many of these processes do occur frequently. These facts make significant evolutionary change (macroevolution) seem inevitable, especially if life has been around for billions of years.
In response to this line of reasoning, some Christians have noted that finding plausible mechanisms by which life could undergo significant evolutionary change doesn’t show that it did, much less that all living things descended from a single-celled ancestor. Many old-earth creationists, and even some young-earth creationists, believe that God supernaturally created just a few original kinds of organisms (perhaps at the taxonomic rank of “order” or “family”) then ordained evolutionary mechanisms to diversify those original kinds into the many species we see today.
Other critics of evolution, and even some advocates of the evolutionary doctrine of universal common descent, have challenged the idea that unguided natural processes can explain the origin of truly novel body structures, organs, and biochemical machinery. For example, biochemist Michael Behe has argued that many of the complicated molecular machines found in living cells are irreducibly complex, meaning that their complexity cannot be reduced without complete loss of function. In other words, a high level of complexity is required for these biochemical machines to function at all. It is difficult to see how undirected processes like random mutation and natural selection could produce such irreducibly complex structures, Behe points out. Since the transitional forms leading up to these biochemical machines would have been useless encumbrances, they should have been eliminated by natural selection.Michael Behe, Darwin’s Black Box: The Biochemical Challenge to Evolution (New York: Free Press, 1996)
Behe’s 1996 book Darwin’s Black Box set off a firestorm of controversy and a series of hasty (often sloppy) attempts to debunk his arguments about irreducible complexity. We’ll consider his main argument, along with some rebuttals to it, in Chapter 11. For now, it is worth mentioning that Behe is a devout Christian in the Catholic tradition, yet he sees no conflict between his faith and mainstream evolutionary science. He accepts the claim that all life descended from a common ancestor. His objection to evolution concerns only the explanatory adequacy of the mechanisms mentioned above.
Behe raised another challenge to the mechanisms of mutation and natural selection in his more recent book Darwin Devolves (2019). There, he argued that those two mechanisms work together to limit evolutionary change. In brief, the problem is that beneficial mutations are usually destructive in the sense that they destroy or degrade functional genes. Since immediate benefits often can be gained by disabling genes that are not needed in the present environment, destructive mutations and natural selection together tend to eliminate functional genes more quickly than constructive mutations can build new ones. Behe calls this the “The First Rule of Adaptive Evolution: Break or blunt any functional gene whose loss would increase the number of a species’s offspring.”Michael J. Behe, Darwin Devolves: The New Science About DNA That Challenges Evolution (New York: HarperCollins, 2019) (Destructive mutations that confer resistance to malaria, mentioned on the previous page, exemplify the rule.) This observation casts further doubt on the claim that increasingly function-rich genomes are the likely result of random mutation and natural selection.
A related issue is the waiting times problem for coordinated mutations. Two or more mutations are said to be coordinated if they must occur together in order to provide an advantage favored by natural selection. The average time required for coordinated mutations to occur can be calculated based on known mutation rates; and, in many cases, the estimated time far exceeds the time available in evolutionary history. The waiting times problem was first recognized not by critics of evolution but by evolutionary scientists as early as 1970, and it has remained an issue of concern in the mainstream scientific literature in recent years.Prominent evolutionary geneticist John Maynard Smith was among the first to anticipate the waiting times problem. In a 1970 letter in the journal Nature, “Natural Selection and the Concept of a Protein Space” (Nature 225: 563–64), he raised the potential problem but quickly dismissed it, predicting that coordinated mutations “are probably too rare to be important in evolution.” (564) Another discussion of the waiting times problem appeared the same year in Bodmer (1970), “The evolutionary significance of recombination in prokaryotes,” Symposium of the Society for General Microbiology 20, 279–294. For more recent examples, see Christiansen, Otto, Bergman, and Feldman (1998), “Waiting with and without Recombination: The Time to Production of a Double Mutant,” Theoretical Population Biology 53(3), 199-215; Schweinsberg (2008), “The waiting time for m mutations,” Electronic Journal of Probability 13, 1442-1478; Durrett, Schmidt, and Schweinsberg (2009), “A waiting time problem arising from the study of multi-stage carcinogenesis,” Annals of Applied Probability 19(2), 676–718; Behrens, Nicaud, and Nicodéme (2012), “An automaton approach for waiting times in DNA evolution,” Journal of Computational Biology 19(5), 550–562; Chatterjee, Pavlogiannis, Adlam, and Nowak (2014), “The time scale of evolutionary innovation,” PLOS Computational Biology 10(9), 1-7; and Hössjer, Bechly, and Gauger (2021), “On the waiting time until coordinated mutations get fixed in regulatory sequences,” Journal of Theoretical Biology 524(110657), 1-37. For example, a 2008 study published in the journal Genetics calculated that a simple two-gene coordinated mutation should take more than 200 million years to occur by chance in human beings.Rick Durrett and Deena Schmidt, “Waiting for Two Mutations: With Applications to Regulatory Sequence Evolution and the Limits of Darwinian Evolution,” Genetics 180, Issue 3 (November 2008), 1501–1509. For an accessible explanation of this calculation and its implications, see Stephen C. Meyer, Darwin’s Doubt: The Explosive Origin of Animal Life and the Case for Intelligent Design (New York: HarperOne, 2013), Chapter 12. That’s as long as the entire evolutionary history of mammals and far longer than human beings—or even primates—have existed.
Similar calculations have led some scientists to doubt that random, unguided mutation is capable of producing novel genes. These doubts are especially pressing because mutation is the only mechanism in mainstream evolutionary theory that could produce new genes and alleles. No other mechanism can do the job. Recombination involves pre-existing alleles, as we have seen; it does not produce new genes or alleles. Likewise, mechanisms that alter the frequencies of traits in a population (natural selection, sexual selection, genetic drift, and gene flow) operate only upon genes that already exist. Even symbiogenesis and the speculative mechanisms proposed in epigenetics and evo-devo have no power to create novel genetic information. So, mutation had better be able to do the work! Can it?
In order to create a new functional gene, mutations must repeatedly shuffle and scramble the nucleobases on a strand of DNA until a sequence is accidentally produced that encodes a functional protein molecule. What are the chances of that happening? To answer this question, it would be helpful to know what proportion of possible nucleobase sequences would yield a functional protein. Beginning in the 1960s, several prominent scientists argued on theoretical grounds that there must be vastly more nonfunctional possibilities than functional ones, making it infeasible for unguided mutations to produce novel genes.Murray Eden, Marcel-Paul Schützenberger, and a number of others raised this problem at a 1966 conference at the Wistar Institute in Philadelphia, where they debated the issue with Ernst Mayr, Richard Lewontin, and other leading evolutionary biologists. In 1985, biochemist Michael Denton pressed the argument further in his book Evolution: A Theory in Crisis (London: Adler and Adler). Denton used an analogy with natural language to suggest that the proportion of functional sequences would likely decrease with sequence length (see especially chapter 13, pp. 308-325). For further discussion of these arguments and their significance, see Stephen C. Meyer, Darwin’s Doubt: The Explosive Origin of Animal Life and the Case for Intelligent Design (New York: HarperOne, 2013), Chapter 9. At the time, however, there was no way to test those claims experimentally.
In 2004, molecular biologist Douglas Axe published the results of a study designed to resolve the issue. While conducting postdoctoral research at Cambridge, Axe had performed a series of mutation experiments to determine how much change a gene could tolerate before it no longer produced a functional protein molecule. Recall that a protein molecule consists of a long chain of amino acids that folds into a specific 3-dimensional shape. Not just any sequence of amino acids will fold itself into a stable structure, however. The vast majority of possible amino acid sequences do not fold into proteins at all, let alone functional ones.
In his experiments, Axe introduced mutations into specific segments of a gene (using a technique called site-directed mutagenesis) and determined what proportion of these mutations destroyed the protein’s ability to fold into a stable structure. For mutations that did produce stable proteins, he performed additional experiments to determine whether the modified protein retained at least some functionality. Then, he extrapolated from those results to estimate the proportion of possible sequences that would yield functional proteins. His conclusion was stunning. Axe found that for every sequence producing a relatively small protein just 150 amino acids long, there are 1074 possible sequences of the same length that don’t fold into proteins at all, let alone functional ones. The proportion of functional sequences was even smaller: about 1 in 1077.Douglas D. Axe, “Estimating the Prevalence of Protein Sequences Adopting Functional Enzyme Folds,” Journal of Molecular Biology 341 (2004), 1295-1315. Axe provides a more accessible explanation of his results in this article. For further discussion of the significance of Axe’s experiment and other similar experiments, see Stephen C. Meyer, Darwin’s Doubt: The Explosive Origin of Animal Life and the Case for Intelligent Design (New York: HarperOne, 2013), Chapter 10; also his book Return of the God Hypothesis: Three Scientific Discoveries That Reveal the Mind Behind the Universe (New York: HarperOne, 2021), Chapter 10, especially footnote 18. That number is about ten trillion, trillion, trillion times greater than the total number of organisms estimated to have ever lived in the entire history of our planet!Recent estimates place the total number of organisms in the history of Earth somewhere in the ballpark of 1040. See Stephen C. Meyer, Darwin’s Doubt: The Explosive Origin of Animal Life and the Case for Intelligent Design (New York: HarperOne, 2013), Chapter 10 (and footnote 19); also Return of the God Hypothesis: Three Scientific Discoveries That Reveal the Mind Behind the Universe (New York: HarperOne, 2021), 205-206. If Axe’ calculations are correct, random mutation is unlikely to have produced even one functional protein molecule of modest length during the entire history of life on earth. This troubling issue has been called the combinatorial problem.
Axe’s conclusions are controversial. One frequent objection is that his calculations must be mistaken, since evolution obviously has produced many new proteins throughout the history of life. This objection simply misses the point: Axe’s results call into question the prevailing scientific account of how new proteins are supposed to evolve. Random, unguided mutations are unlikely to yield novel proteins. Moreover, contrary to a common misconception, natural selection doesn’t help to resolve the combinatorial problem. Natural selection may optimize the functionality of a protein that already exists, but it can’t guide mutations to produce a novel protein fold in the first place. Even worse, natural selection acts to prevent an already-optimized protein from evolving into a different protein.For further explanation and discussion of these points, see Stephen C. Meyer, Darwin’s Doubt: The Explosive Origin of Animal Life and the Case for Intelligent Design (New York: HarperOne, 2013), Chapter 10.
A more plausible response to the combinatorial problem alleges that the specific protein used in Axe’s experiments may be atypical. Perhaps, if he had done the experiment with other proteins instead, the results would have been different. Some studies involving other proteins have indeed yielded higher estimates of the proportion of functional proteins among possible amino acid sequences.Evolutionary creationist Deborah Haarsma raises this objection to Axe’s results, citing two such studies: this and this. She also cites two blog posts by Dennis Venema that have since been retracted. Venema’s now-discredited argument had involved the nylon digesting enzyme nylonase, which I mentioned in a fineprint section on the previous page. See Haarsma, “Response from Evolutionary Creation,” in J.B. Stump and Stanley N. Gundry (eds.), Four Views on Creation, Evolution, and Intelligent Design (Grand Rapids: Zondervan, 2017), 224. On the other hand, even if the protein Axe studied was atypical, the puzzle remains: how could random, unguided mutations produce proteins like that one, which requires such exquisite precision in its amino acid sequence before it can fold into a stable (and potentially functional) structure? Regardless of whether Axe’s results are typical of proteins in general, his calculations pose a serious challenge to the idea that all novel proteins were produced by random mutations.