Evolution Unscathed: A Review of Darwin Devolves

[swamidass (S. Joshua Swamidass)]

https://onlinelibrary.wiley.com/doi/10.1111/evo.13710

Evolution unscathed: Darwin Devolves argues on weak reasoning that unguided evolution is a destructive force, incapable of innovation

Authored by two of Behe’s colleagues at Lehigh University’s Department of Biological Sciences,

Academic buildings can be quirky places. Like proteins they evolve, adapting to new functions while holding on to traces of their past, occasionally performing tasks that are in conflict with one another. Iacocca Hall is no exception. Once home to the research and development division of Bethlehem Steel, the mountaintop building now houses Lehigh University’s Department of Biological Sciences. The most infamous (among scientists) and celebrated (among creationists) of our 22 faculty is Michael Behe, a well‐known proponent of intelligent design and a tenured member of our department who recently released his third book, Darwin Devolves: The New Science About DNA That Challenges Evolution (HarperOne, 2019, 352 pp.).

This is a phenomenal article. The best of the reviews that have come out, echoing, expanding, and adding too much of what the rest of us have said. Some greate quotes:

About Real Time Evolution

Proteins are promiscuous (Hammer et al. 2017). They moonlight, by chance interacting with other cellular components to effect phenotype outside their traditionally ascribed roles (Copley 2014). These adventitious functions can be strengthened by selection, allowing a protein to assume a new or a dual role. This topic was raised in a recent review of Darwin Devolves in Science (Lents et al. 2019), where the authors highlight a study that employed experimental evolution to strengthen the weak nascent ability of a protein in the histidine biosynthesis pathway to act on a similar substrate in tryptophan biosynthesis (Näsvall et al. 2012). For multifunctional proteins, gene duplication and divergence can parse specific functions into separate proteins, each now free to specialize to its own task (Hittinger and Carroll 2007).

About Misleading Metaphors

The planthopper’s hind legs are a “large, in‐your‐face, interacting gear system (p45).” As for the “supremely complex molecular machine called the spliceosome,” simple scissors will not do. Behe opts for “an automated machine that could make fancy paper cut‐out dolls (p61).” Most of the analogies in Darwin Devolves are not Behe’s creation—he has done well to scour press coverage (Yong 2013) and the scientific literature (Burrows and Sutton 2013; Franze et al. 2007) for relatable metaphors; and he is generous with their use. But reality remains: proteins are not machines, a flagellum is not an outboard motor, and no one is claiming that mousetraps evolved gradually by natural selection.

Behe nearly catches his own error. He notes: “Metaphors, of course, do not denote the thing they are applied to. They are vague analogies… [I]f metaphors are taken too seriously in science, they can be confusing at best; and they can often be actively misleading (p201).”

Much more to dig into here.

[swamidass (S. Joshua Swamidass)]

I would very much agree with this:

Largely funded by federal grants, scientists have a moral responsibility (if not a financial obligation) to ensure that the core concepts of our respective fields are communicated effectively and accurately to the public and to our trainees.

[T_aquaticus]

As an aside, and in the name of finding agreement, the review linked in the OP noted the following:

In his previous book, The Edge of Evolution (Behe 2007), he writes that “distantly related organisms share apparently arbitrary features of their genes that seem to have no explanation other than that they were inherited from a distant common ancestor (p3).” Regarding our closest cousins, “[i]t’s hard to imagine how there could be stronger evidence for the common ancestry of chimps and humans (p72).”

[T_aquaticus]

Citing sections I find important as I go:

If Behe is trying to claim that gene loss will result in all genes being lost, then good ol’ brewers yeast (whose hard work I will be enjoying this evening) demonstrates otherwise. Gene loss matches gene gain.

For example, along the lineage leading to modern brewers’ yeast ( Saccharomyces cerevisiae ), historical rates of gene loss and gene gain are nearly matched such that the inferred ancestral genomes along the way are all predicted to have around 6000 genes (Wapinski et al. 2007), despite a large flux in gene content (including a whole‐genome duplication event;Wolfe and Shields 1997; Kellis et al. 2004).

One aspect many of us have not focused on is the importance of regulatory elements in evolution.

When marine sticklebacks independently colonized multiple freshwater glacial lakes, an allele that was segregating at low frequency in ancestral marine stickleback populations, the low‐ Eda allele, was favored by natural selection and rose to high frequency in the parallel lake populations, leading to a beneficial reduction in armored plating (Colosimo et al. 2005). Counter to Behe’s thesis, none of the observed Eda mutations were predicted to damage the function of the protein (Colosimo et al. 2005). The causative variants are likely cis ‐regulatory changes that decreased expression of Eda in developing armor, but not in other tissues (O’Brown et al. 2015). Darwin Devolves accepts as evidence only de novo protein evolution, a restriction Behe uses to support his “First Rule” and claim that “Darwinian evolution is self‐limiting.” Narrow by definition and unsupported by the data, Behe’s First Rule does not stand up to scrutiny.

Great example of proteins being IC in some populations but not in others. This demonstrates how IC systems evolve.

Simply because a system in its current form is irreducibly complex is not evidence that it did not evolve by random mutation and natural selection. Essentiality of a gene or protein is relative to its current state. For two closely related strains of yeast, between 1% and 5% of genes that are essential in one strain are dispensable in the other (Dowell et al. 2010).

Example of an IC system evolving.

Systems that meet Behe’s definition of irreducible complexity do, in fact, evolve by random unguided processes. In most eukaryotes, the vacuolar ATPase contains a six‐member V0 ring consisting of five subunits of Vma3 and one subunit of Vma16. In the ancestor of all fungi the VMA3 gene was duplicated around the time that fungi split from all other eukaryotes, approximately one billion years ago (Finnigan et al. 2012). Initially Vma3 and its paralog, called Vma11, were interchangeable; however, over time the binding interfaces of Vma3 and Vma11 degraded such that Vma11, Vma3, and Vma16 all became essential components of the vacuolar ATPase (Finnigan et al. 2012). The increase in the complexity of the vacuolar ATPase is a clear example of how gene duplication and degenerative mutations—processes Behe admits occur by random unguided processes—can result in an increase in complexity and molecular novelty.

The reviewers had the same view of Behe’s definitions of IC and the presence of IC in the evolution of lambda phage.

Ever the rigid selectionist, Behe has offered an alternative definition of irreducible complexity to contend with evolutionary scenarios like the vacuolar ATPase: “An irreducibly complex evolutionary pathway is one that contains one or more unselected steps (that is, one or more necessary‐but‐unselected mutations) (Behe 2000).” . . .

Irreducible complexity also fails to account for the reality that most proteins—unlike the fishing rods, hammers, and gears in Darwin Devolves —perform multiple functions, bind to multiple other proteins, and thus evolve on multiple fitness landscapes simultaneously. Experimental evolution of lambda bacteriophage in coculture with E. coli showed that phage can evolve the ability to bind a novel receptor, OmpF (Meyer et al. 2012). Each of four mutations in the phage J protein (a tail protein that mediates host binding) is required in an “all‐or‐none” manner in order for the J protein to bind to OmpF. This evolutionary pathway satisfies Behe’s definition of irreducible complexity: remove any one of these four interacting mutations and binding to OmpF is abolished. How then did these four mutations arise? It turns out that individually, these mutations improve binding to the original receptor LamB. While adapting along a step‐wise evolutionary path of improved binding to LamB, the J protein was able to acquire the ability to bind to a new receptor—to evolve a novel protein–protein interaction that was dependent upon four mutations acting in an all‐or‐none manner, well beyond Behe’s “edge of evolution.”

All of these could be the start of a thread all on their own. The review article is what the military would call a “target rich environment”.

[swamidass (S. Joshua Swamidass)]

@T_aquaticus

Initially Vma3 and its paralog, called Vma11, were interchangeable; however, over time the binding interfaces of Vma3 and Vma11 degraded such that Vma11, Vma3, and Vma16 all became essential components of the vacuolar ATPase (Finnigan et al. 2012). The increase in the complexity of the vacuolar ATPase is a clear example of how gene duplication and degenerative mutations—processes Behe admits occur by random unguided processes—can result in an increase in complexity and molecular novelty.

That is exactly right, and I have been planning to make this clear. By steps that each on their own are “devolution,” we can evolve irreducible complexity. We can “devolve” into irreducible complexity.

[NLENTS (Nathan H. Lents)]

But please don’t keep using the words “devolve” and “devolution.” If these terms sneak into our language, this will only breed confusion and serve to undercut the hard work we’re doing to clarify evolutionary theory for the public.

[swamidass (S. Joshua Swamidass)]

Hence forth, the term “devolve” will be comfortably cushioned within quotes, as an indicator we are using an idiosyncratic term outside established science.

[NLENTS (Nathan H. Lents)]

and in spoken form, air/finger quotes must be used each and every time. It is so ordered.

[Mark]

@T_aquaticus

Systems that meet Behe’s definition of irreducible complexity do, in fact, evolve by random unguided processes.

I think this is why so many Christians gravitate toward ID. These terms need to be defined. If random is defined as every outcome being equally likely, this is not true. Behe should use “external intelligence” rather then “unguided.”

When responded to, this review makes it sound like “Behe is wrong. Evolution IS purposeless and an atheistic process.” Off goes the Christian to ID.

What if they said “IC is produced through several natural processes without external intelligent input through changes that are highly constrained due to physics and chemistry.”

Sounds a lot less threatening to me than “random and unguided processes.”

I don’t want to go into a discussion of the definition of random, but…yeah. Behe AND the reviewers need to precisely define their terms.

[T_aquaticus]

@Mark

I think this is why so many Christians gravitate toward ID. These terms need to be defined. If random is defined as every outcome being equally likely, this is not true.

In this case, mutations are random with respect to fitness. It’s a bit like the game of craps. Not all outcomes are equal. 2 and 12 are relatively rare while 7 is the most common outcome. The important part is that the dice aren’t influenced by the chips on the table. Putting your bet on a hard 8 doesn’t increase the probability of rolling a hard 8. The dice are blind to the chips. In the same way, the processes that cause mutations are blind to what the organism needs in a given environment, or at least they appear to be blind by all of the statistical tests that have been used.

This gets us to the other important bit. Science uses statistics to define randomness. Science doesn’t make any claims about God’s involvement, or more specifically, about God’s involvement in mutations. Could God be guiding everything from the roll of the dice to mutations in a way that we can’t perceive using our statistical tools? Sure, I guess so. However, science is silent on that topic.

@Mark

What if they said “IC is produced through several natural processes without external intelligent input through changes that are highly constrained due to physics and chemistry.”

What if they just said, “IC is produced through several natural processes”. I think that is all that needs to be said. At the same time, ID proponents seem to dislike the idea of natural processes being capable of producing what we see in biology. They seem to require supernatural input outside of the natural processes we observe.

@Mark

I don’t want to go into a discussion of the definition of random, but…yeah. Behe AND the reviewers need to precisely define their terms.

The reviewers are going to be using the standard definition of random that I described above. I may have violated your wishes, but hopefully it can supply some insight into how scientists define randomness.

[swamidass (S. Joshua Swamidass)]

@T_aquaticus

In this case, mutations are random with respect to fitness.

Please stop saying this. We discussed several times this is not precisely true

[T_aquaticus]

@swamidass

Please stop saying this. We discussed several times this is not precisely true

There are very few, if any, concepts in biology that don’t have exceptions.