Biologic Institute

The second edition of The Design Inference was recently released. Winston Ewert from the Biologic Institute coauthored the book with William Dembski who wrote the first edition. The original book presented a general methodology for detecting design, which was the most sophisticated framework to date. The second edition represents the culmination of two decades of responding to critics and refining his model. Princeton University mathematician Sergiu Klainerman welcomed the book as follows:

Well argued and eminently readable… I don’t see how any open-minded scientist can ignore this important book.

The first edition of The Design Inference explained how design detection involves identifying patterns that are both highly improbable and specified. A common question raised by readers was the precise definitions for the terms specified and specification. In other words, what exactly makes a pattern specified or special? In many contexts, the answer is simple. For instance, being dealt four poker hands that were all royal flushes would be a very rare and special pattern pointing to someone cheating. Design detection in this example is easy because the category royal flush was already defined (aka prespecified) by the rules of the game. The situation is more challenging when categories of outcomes are not so clearly defined. 

In the new edition, Dembski and Ewert greatly refined the concept of specifications. They defined a specification as a description that requires few words. The description must also only apply to a small proportion of possible outcomes, so the probability of any outcome falling within the specification must be sufficiently small. Patterns that are both identifiable by a short description and highly improbable are almost always designed.  

Returning to the poker example, the four hands I mentioned could be described or specified by the three words four royal flushes, and the probability of four random hands falling within this specification is less than 1 chance in 100 billion trillion. The short description and very low probability indicate design. 

In contrast, four typical hands of five cards would require a very long description, such as the following:

two of diamonds, pair of threes, pair of jacks; three fours, nine of hearts, ten of clubs; four of diamonds, pair of eights, nine of clubs, queen of clubs; three kings, pair of aces

The probability of four hands falling within this description is low, but the description length is very long. Many sets of four poker hands would fall within a description of that length. The low probability is offset by the large number of sets of four hands with descriptions that are that long. The probability of being dealt four hands with such a long description is not sufficiently low to conclude design. 

Conversely, one could label four generic hands with the description random cards. The description is now short, but the probability is very high of four hands falling within this description. The short description is offset by the high probability. Similarly, the famous face on Mars can be described by the two words fuzzy face. Here again, the description length is short, but the probability of a photograph of the surface of a celestial body resembling a face at least as well as the Mars photo is sufficiently high to occur by chance. 

The face on Mars (NASA/JPL)

The underlying logic of Dembski and Ewert’s methodology is profound. Identifying design requires a pattern being assigned by a mind a special significance. A short description meets this criterion since societies assign words or short phrases to that which they designate as special. 

To illustrate, in the movie The Empire Strikes Back, Darth Vader declares to Luke Skywalker one of the most iconic lines in movie history: "I am your father."

Darth Vader could specify himself by the single word father. The relationship between a father and son is designated by all societies as a very special relationship. Relatively few other people could be specified by such a short description. Consequently, the audience knew Luke’s meeting Vader was not an accident but by design. 

In stark contrast, Dark Helmet in the movie Spaceballs tells Lonestar a far longer statement: "I am your father’s brother’s nephew’s cousin’s former roommate."

Dennis Venema, associate professor of biology at Trinity Western University, and Scot McKnight, professor of New Testament at Northern Seminary, have written a book, Adam and the Genome: Reading Scripture after Genetic Science, that addresses the question of human origins. Venema defends the standard scientific narrative on neo-Darwinism and human origins, while McKnight attempts to reconcile Christian theology with the science. Because the book argues that science disproves the existence of a historical Adam and Eve, it is worth responding to in some detail.

One of the claims Venema makes in his book is that the effective population size of our last common ancestor with chimps has never been fewer than 10,000. In fact, he equates the certainty of that statement with our certainty about heliocentrism. That is an extreme claim that needs justification. Assuming no bottlenecks, this would exclude the possibility of an original human pair as the progenitors of the human race.

Effective population size is very hard to determine. In fact, in some situations the effective population size cannot be determined (On the Meaning and Existence of an Effective Population Size P. Sjödin, I. Kaj, S. Krone, M. Lascoux and M. Nordborg (2005) Genetics 169: 1061–1070). So any estimates concerning effective population size should be taken a bit skeptically.

You don’t have to take my word for it. Richard Buggs, a British biologist who has published over thirty articles on genetics in journals such as Nature, Current Biology, Evolution, and Philosophical Transactions of the Royal Society, wrote to Venema to lay out his concerns in May of this year. Venema failed to respond to Buggs’s email so Buggs posted it online last week, and tweeted a link to it here. In his letter he comments on effective population size estimates:

As I am sure you know, effective population size is a measure of a population’s susceptibility to drift, rather than an attempt to measure census population size. I would be very hesitant to rely too heavily on any estimate of past effective population size.

Could we have begun with a population of 10,000 and then had a bottleneck of two? Obviously this estimate is relevant for human origins, since a bottleneck of two would indicate some sort of unique beginning. Venema says a bottleneck of two is impossible based on the levels of human genetic diversity present today.

If a species were formed through such an event [by a single ancestral breeding pair] or if a species were reduced in numbers to a single breeding pair at some point in its history, it would leave a telltale mark on its genome that would persist for hundreds of thousands of years — a severe reduction in genetic variability for the species as a whole.

Buggs responds:

It is easy to have misleading intuitions about the population genetic effects of a short, sudden bottleneck. For example, Ernst Mayr suggested that many species had passed through extreme bottlenecks in founder events. He argued that extreme loss of diversity in such events would promote evolutionary change. His intuition about loss of diversity in bottlenecks was wrong, though, and his argument lost much of its force when population geneticists (M. Nei, T. Maruyama and R. Chakraborty 1975 Evolution, 29(1):1-10) showed that even a bottleneck of a single pair would not lead to massive decreases in genetic diversity, if followed by rapid population growth. When two individuals are taken at random from an existing large population, they will on average carry 75% of its heterozygosity (M. Slatkin and L. Excoffier 2012 Genetics 191:171–181). From a bottleneck of a single fertilised female, if population size doubles every generation, after many generations the population will have over half of the heterozygosity of the population before the bottleneck (Barton and Charlesworth 1984, Ann. Rev. Ecol. Syst. 15:133-64). If population growth is faster than this, the proportion of heterozygosity maintained will be higher.

This means that a single pair of individuals can carry a great deal of heterozygosity with them through a bottleneck, provided they come from an ancestral population with high diversity and undergo rapid population growth after the bottleneck. They will pass most of that diversity on to the population they found, so long as population grows rapidly.

Buggs goes on to discuss the relative genetic diversity of chimpanzees at 5.7 million SNPs (single nucleotide polymorphisms) and humans at 3.1 million (Prado-Martinez et al 2013 Nature). He makes the point that if a pair of chimpanzees were moved to an isolated region where rapid population growth was possible, the new population would have similar levels of genetic variability to modern humans.

He explains:

I am not stating these figures because existing populations of chimpanzee gave rise to modern humans, but simply to show that it is hard to see how overall levels of SNP diversity and heterozygosity in modern humans could exclude the possibility of a past bottleneck of two individuals.

On top of this, we need to add in the fact that explosive population growth in humans has allowed many new mutations to rapidly accumulate in human populations, accounting for many SNPs with low minor allele frequencies (A. Keinan and A. G. Clark (2012) Science 336 (6082): 740-743).

PSMC (pairwise sequentially Markovian coalescent) analysis is a method used by population geneticists to try to recover the history of human populations. Venema discusses it in his book. Unfortunately he cannot use it to prove that a sudden, short bottleneck never happened. In the original Li and Durbin 2011 paper describing PSMC analysis, the authors note that an instantaneous reduction in population size was not detected as such, but was instead “spread over several preceding tens of thousands of years in the PSMC reconstruction.” Continued work in Beth Shapiro’s lab also indicates that the PSMC method cannot accurately reconstruct sharp bottlenecks.

Buggs sums up his problems with Venema’s analysis:

In general, I am concerned that the studies you cite did not set out to test the hypothesis that humans have passed through a single-couple bottleneck. They are simply trying to reconstruct the most probable past effective population sizes of humans given the standard assumptions of population genetic models. I personally would feel ill at ease claiming that they prove that a short sudden bottleneck is impossible. [Emphasis added.]

Other scientists have argued against Venema’s position by disputing key assumptions, including the idea of common descent. Buggs’s argument is valid, though, whether or not common descent is true. In fact, he has told me he is assuming common ancestry, but arguing that we can still have come from a single couple as ancestors.

What is needed is a model that does not rely on the standard assumptions of population genetics. Recently Hössjer, Gauger, and Reeves have proposed such an alternative model. The model, when programmed, will be available for use by anyone to test the effects of various starting conditions, like population size, the degree of gene flow, recombination and mutation rates, mate choice effects, and ultimately the effects of selection. The outcomes of various scenarios can be tested and compared to summary statistics from modern populations, in order to determine which scenarios best explain current populations.

Finally, given Buggs’s critique, Venema’s claims that our starting population was at least 10,000, and that we could not have come from just two, seem unjustified. Certainly Venema cannot claim these are facts as sure as heliocentrism. He should at least acknowledge the facts above, and soften his claims. Further revision may be necessary following the results from Hössjer, Gauger, and Reeves’s model. We shall see.

Photo credit: music4life, via Pixabay.

Nylonase: Move On, Nothing to See Here, says Theistic Evolutionist

Ann Gauger

The supposed sudden emergence of the enzyme nylonase has been a chief talking point for the power of evolution for many years, and it has made its appearance multiple times in Dennis Venema’s series of posts, “Letters to the Duchess,” at the theistic evolutionist website BioLogos. Here’s how Venema, Fellow of Biology, described it in a post last year:

Intelligent Design and Nylon-eating Bacteria

[Nylonase] arose from scratch as an insertion mutation into the coding sequence of another gene. This insertion simultaneously formed a “stop” codon early in the original gene (a codon that tells the ribosome to stop adding amino acids to a protein) and formed a brand new “start” codon in a different reading frame. The new reading frame ran for 392 amino acids before the first “stop” codon, producing a large, novel protein. As in our example above, this new protein was based on different codons due to the frameshift. It was truly “de novo” — a new sequence.

Why was this supposed to be important? Venema continued:

In order for proteins to function, they need to fold up into stable shapes — “protein folds”. Meyer claims, based on the work of his colleague Douglas Axe, that stable protein folds are vanishingly rare — on the order of only one in 10 to the 77th power. As such, Meyer argues, evolution cannot find the needle (a functional protein) in the haystack (the vast number of functionless possibilities) … All of this hinges, of course, on just how accurate Axe’s estimate is: are functional proteins really that rare?…

So if nylonase arose by frameshift mutation, via a single base insertion, producing a novel sequence and a stable, functional fold, it would appear that functional, stable folded proteins are not rare. Venema said:

…this brand-new protein [nylonase] folded into a stable shape, and acted as a weak nylonase. Later duplications, mutations and selection would make a very efficient nylonase from this starting point. Additionally, the three-dimensional structure of the protein has been solved using X-ray crystallography, a method that gives us the precise shape of the protein at high resolution. Nylonase is chock full of protein folds — exactly the sort of folds Meyer claims must be the result of design because evolution could not have produced them even with all the time since the origin of life.

Put another way, if only one in 10 to the 77th proteins are functional, there should be no way that this sort of thing could happen in billions and billions of years, let alone 40. Either this was a stupendous fluke (and stupendous isn’t nearly strong enough of a word), or evolution is in fact capable of generating the information required to form new protein folds.

And if this can happen in 40 years, what might millions of years of evolution produce? [Emphasis added.]

Venema was merely repeating the old story that has long circulated around many anti-ID websites. It goes back to 1984, when Susumu Ohno first proposed a frameshift mutation as the cause of the new nylonase enzyme.

Venema made nylonase a key element in his argument against intelligent design. He repeated the story in another post this past January (“Biological Information and Intelligent Design: evolving new protein folds”), and in his recent book, Adam and the Genome, where he said:

…what would truly settle the issue is a concrete example of a “truly new” gene coming into being — one that is not a duplicate of a preexisting protein sequence — that nonetheless has a new function and new protein folds. If scientists could observe such an event, then it would indicate that Axe’s math (and Meyer’s use of it) is not a reliable estimate for the prevalence of functional protein folds.

Now, here’s where I come in. It seemed highly unlikely to me that a stable, functional protein fold could arise by frameshift mutation, unless the sequence was designed that way. Among geneticists frameshift mutations are called “nonsense” mutations because they typically result in scrambled non-functional code. In fact, this view of frameshift mutations was pretty universal until we began to encounter what looked like frameshifts in sequenced genomes.

So I began to dig in the literature. What I found was a major block across the road of the old nylonase story. To get the full picture, I’d suggest reading this post I wrote, “The Nylonase Story: When Imagination and Facts Collide,” back in May. It is based on the paper by Negoro et al. describing nylonase’s structure. The paper makes clear that nylonase was a variant of a pre-existing β-lactamase fold, not a novel protein fold. In fact, among the different kinds of β-lactamases, nylonase most resembled carboxylesterases. It was then found that nylonase had previously undetected carboxylesterase activity.

In other words, nylonase had started out as a carboxylesterase, then evolved the ability to degrade nylon byproducts. The title of the paper, and its Abstract, should make the truth obvious:

X-ray crystallographic analysis of 6-aminohexanoate-dimer hydrolase: molecular basis for the birth of a nylon oligomer-degrading enzyme.

Here, we propose that amino acid replacements in the catalytic cleft of a preexisting esterase with the β-lactamase fold resulted in the evolution of the nylon oligomer hydrolase [nylonase].

Nylonase did not have a frameshifted new protein fold. It was a pre-existing fold with activity characteristic of its fold type. No novel protein fold had emerged. And no frameshift mutation was required to produce nylonase. To quote my post:

Where did the nylon-eating ability come from? Carboxylesterases are enzymes with broad substrate specificities; they can carry out a variety of reactions. Their binding pocket is large and can accommodate a lot of different substrates. They are “promiscuous” enzymes, in other words. Furthermore, the carboxylesterase reaction hydrolyzes a chemical bond similar to the one hydrolyzed by nylonase.

From Kato et al. (1991):

“Our studies demonstrated that among the 47 amino acids altered between the EII and EII’ proteins, a single amino acid substitution at position 181 was essential for the activity of 6-aminohexanoate-dimer hydrolase [nylonase] and substitution at position 266 enhanced the effect.”

So. This is not the story of a highly improbable frame-shift producing a new functional enzyme. This is the story of a pre-existing enzyme with a low level of promiscuous nylonase activity, which improved its activity toward nylon by first one, then another selectable mutation. In other words this is a completely plausible case of gene duplication, mutation, and selection operating on a pre-existing enzyme to improve a pre-existing low-level activity, exactly the kind of event that Meyer and Axe specifically acknowledge as a possibility, given the time and probabilistic resources available. Indeed, the origin of nylonase actually provides a nice example of the optimization of a pre-existing fold’s function, not the innovation or creation of a novel fold.

The data described in the Negoro et al. and Kato et al. papers build a brick wall right across the path of Venema’s argument. And sadly, that roadblock pre-existed Venema’s posts. If he had looked carefully he would have seen it.

What do you suppose Venema’s reaction was to this collision of imagination and fact? He responded by saying, essentially, “Move on, nothing to see here.” A combination of selective retelling, sleight of hand, redirection, and downgrading the importance of the nylonase story were his main techniques. Let’s take Venema’s final post on nylonase a bit at a time to see how he did it. Dated May 18, 2017, it bears the headline:

Biological Information and Intelligent Design: New functions are everywhere

He begins with the canonical story, simply restated. No mention of the contrary facts.

In previous posts in this series, we’ve explored the claim made by the Intelligent Design (ID) movement that evolutionary mechanisms are not capable of generating the information-rich sequences in genes. One example that we have explored is nylonase — an enzyme that allows the bacteria that have it to digest the human-made chemical nylon, and use it as a food source. As we have seen, nylonase is a good example of a de novo gene — a gene that arose suddenly and came under natural selection because of its new and advantageous function.  Since nylonase is a folded protein with a demonstrable function, it should be beyond the ability of evolution to produce, according to ID.

But Venema did notice I was writing about the subject. I assume he read what I wrote.

The implications of nylonase for ID arguments are clear, and they have caught the notice of several ID supporters. In recent weeks a number of posts on the subject have appeared on ID websites. ID biologist Anne [sic] Gauger, for example, is writing a series of posts in an attempt to rebut the evidence that nylonase is in fact a de novo gene. Her motivation for this work is clear. [Emphasis added.]

But at this point, oddly enough, rather than describe and refute my arguments, he changes the subject. He spends the rest of the post discussing a new paper that deals with other data regarding the ease with which random sequences benefit bacteria. He only mentions nylonase to dismiss it, in the closing paragraphs.

As we can see, the issue at hand is not so much the specific origins of nylonase, but rather the relative ease by which new, functional proteins can come into existence. If it is easy to evolve them, ID advocates are wrong. If new protein functions are vanishingly rare and inaccessible to evolution, ID would be strongly supported. With nylonase, we are dealing with events that happened in the past, so our inferences are limited to working with the evidence we have in the present.

So, we can see that the nylonase issue is something of a distraction — a missing of the forest to focus on one particular tree. Even if this particular example could have an alternate explanation, as Gauger argues, the problems for ID do not go away. [Emphasis added.]

This is a classic tactic. When shown to be wrong about nylonase, Venema changes the subject. Now nylonase is something that happened in the past, he says, implying we can’t reliably make inferences about it. He even says that nylonase is a “distraction” — despite his own repeated use of the nylonase story in trying to argue against Stephen Meyer and Doug Axe.

So as to acknowledge what Venema did talk about, I offer a few comments about the issue of functional, random sequences. Venema described the results of a recent study where a verysurprising number of random sequences were reported to confer a fitness advantage on cells expressing them. I have my doubts about these results and will address them on another occasion. But I am not the only one who should be having doubts. Whatever those proteins or RNAs are doing, it is unlikely to be via a stable functional fold.

François Jacob wrote a famous essay, “Evolution and Tinkering,” in which he said, speaking of the time when the first primitive cells had emerged:

New functions developed as new proteins appeared. But these were merely variations on previous themes. A sequence of a thousand nucleotides codes for a medium-sized protein. The probability that a functional protein would appear de novo by random association of amino acids is practically zero. In organisms as complex and integrated as those that were already living a long time ago, creation of entirely new nucleotide sequences could not be of any importance in the production of new information. [Emphasis added.]

Ann Gauger

Editor’s note: Nylon is a modern synthetic product used in the manufacturing, most familiarly, of ladies’ stockings but also a range of other goods, from rope to parachutes to auto tires. Nylonase is a popular evolutionary icon, brandished by theistic evolutionist Dennis Venema among others. In a series of three posts, of which this is the third, Discovery Institute biologist Ann Gauger takes a closer look. Look here for the first and second posts.

Returning to the story of the nylonase gene and the problem of where new information comes from, I’d like to make the point that there is a reason that molecular geneticist and evolutionary biologist Susumu Ohno made his hypothesis about a frame-shift having produced nylonase. Ohno is famous for his hypothesis that gene duplication and recruitment are the chief means by which “new” proteins are made — he wrote a famous book about it.

But he also knew that copying and tinkering weren’t enough, that there had to be a way to generate genuine de novo information, brand new coding sequence for genuinely new proteins, in order to account for all the diversity of information that must have been necessary as life became more complex. New proteins had to come from somewhere.

Ohno had an idea. He thought coding sequences made up of oligomeric repeats might allow there to be several alternate ways to read the same sequence. For an explanation of alternate reading frames, see my earlier post, “The Nylonase Story: How Unusual Is That?”

As a potential example, Ohno proposed nylB, the gene for nylonase. This gene has certain characteristics that make it plausible that a frameshift could have occurred, characteristics I described in that second post in this series, such as nylB’s sequence being GC-rich and deficient in TAs. These two characteristics reduce the chances of having stop codons, in any frame.

Ohno thus proposed that nylonase arose after a frameshift mutation in a perhaps nonfunctional, prior-coding sequence, resulting in an entirely new coding sequence with nylonase activity. The only reason Ohno could make this proposal was because nylB, the gene that codes for nylonase, has at least two potential open reading frames in the forward direction — the hypothetical “original” one proposed by Ohno from before any hypothetical T insertion took place, and the actual one that codes for nylonase now.

Ohno published his paper in 1984. In 1992, Yomo et al. noticed that one frame in the antisense direction of nylonase has no stop codons either. It also lacks a start codon, though, so Yomo et al. called it a non-stop frame (NSF) instead of an open reading frame (ORF). The probability of finding a DNA sequence with an ORF on the sense strand and a full NSF on the antisense strand are small. But surprisingly, not only does nylB have an NSF on the antisense stand, nylBhas another fully overlapping NSF in the forward direction. That’s two NSFs plus the actual ORF for nylonase (I’m not counting the hypothetical frame-shifted “original” ORF, since that frame actually has several intervening stops. (See “The Nylonase Story: When Facts and Imagination Collide.”) That means nylB has no stop codons in three out of six frames.

The chances of avoiding a stop codon in three out of six frames are very low. Our simulation (described in the previous post) showed that the probability is very small indeed. For an ORF 900 nucleotides long to have two NSFs at 70 percent GC is 9 out of 28,603 or .0003. (See there for details.) If these figures are recast to include the total number of random trials required to get an ORF of the proper length and GC content in the first place, and then with two NSFs, then the probability would be nine out of ten million trials, or 0.0000009. No organisms have ten million genes (we only have about twenty thousand), and Flavobacterium certainly doesn’t. But it’s not outside the realm of possibility that such sequences should exist by pure chance somewhere. After all, nylB does. But take the following into consideration.

In addition, beyond the first appearance of such a sequence, there would also need to be some way to prevent random mutation from introducing any stop codons over evolutionary time, in any of the three open frames. Purifying selection would normally be invoked in such a case. Organisms that develop harmful mutations in genes that encode functional gene products — things that are important for the organism’s survival — are less successful at reproducing, and so organisms carrying harmful mutations tend to disappear from the population (they are sickly or dead). However, purifying selection by definition has no effect on non-functional sequences. The fact that stops are prevented from accumulating in nylB NSFs implies that all three frames are functional. No function has been reported for the NSFs, however. They have no ATGs in the vicinity and so may be non-coding (though it must be acknowledged there are alternate start codons in the vicinity). In addition, it has been reported that the pOAD2 plasmid on which nylB is located is non-essential. It can be cleared from its host with no effect, except the loss of the ability to degrade nylon.

One possibility is that nylB has a secondary DNA or RNA-based function that requires its sequence to be nearly completely conserved. It would have to be a very specific sequence requirement to prevent the accumulation of stop codons in three frames, though. We get a hint that the cause is not sequence specificity, because the nylB and nylB′ genes of Flavobacterium differ by 47 amino acids, and the nylBgene of Pseudomonas has only about 35 percent identity according to reports, yet all three lack stops in the anti-sense frames in addition to their coding sequence (based on available sequence information).

Yomo et al., who first reported the anti-sense NSF in nylB, were amazed and puzzled by the existence of anti-sense NSFs in nylB genes of multiple species.

The probability of the presence of these NSFs on the antisense strand of a gene is very small (0.0001-0.0018) [we observed .0001]. In addition, another gene for nylon oligomer degradation [Pseudomonas nylB] was found to have a NSF on its antisense strand, and this gene is phylogenetically independent of the [Flavobacterium] nylB genes. Therefore, the presence of these NSFs is very rare and improbable. Even if the common ancestral gene of the nylB family was originally endowed with an NSF on its antisense strand, the probability of this original NSF persisting in one of its descendants of today is only 0.007. Unless an unknown force was maintaining the NSF, it would have quickly disappeared by random emergences of chain terminators. Therefore, the presence of such rare NSFs on all three antisense strands of the [three member] nylB gene family suggests that there is some special mechanism for protecting these NSFs from mutations that generate the stop codons. Such a mechanism may enable NSFs to evolve into new functional genes and hence seems to be a basic mechanism for the birth of new enzymes. [Emphasis added.]

Later on, they continue:

… the lifetime of a nonessential NSF is very short, and it is impossible for such a NSF to persist for a long period of evolution. Therefore, we strongly suggest that the existence of the NSFs on all the three antisense strands of the nylB gene family points to an unknown force that is preserving these nonessential NSFs; otherwise, they would have quickly disappeared by random emergences of chain terminators.

Ohno himself was aware of this work and in some sense supported it. He was the one who communicated it to the Proceedings of the National Academy of Sciences. What he made of it I don’t know.

The highlighted proposal in the above quotes is on the face of it antithetical to the materialist worldview. What kind of force can preserve apparently non-functional NSFs? Certainly a mechanism to preserve non-functional sequences so that they might some day evolve into functional genes is more suggestive of design than evolution. It would take a fair amount of foresight on the part of evolution, don’t you think, to develop a mechanism to prevent stop codons from interrupting non-functional NSFs, all for some possible future benefit?

All this speaks to the origin and preservation of potential information, information such as Ohno was looking for, but by a means different than he foresaw. We have returned full circle.  Explaining nylonase does not require a frameshift, as I have shown in the first post — nonetheless nylonase’s gene is an unusual sequence. Getting overlapping code in three frames might happen in very rare circumstances, but keeping the NSFs open in the apparent absence of selection to maintain them would seem to be highly, highly unlikely. So we have extreme rarity piled upon rarity. Bear in mind also, that whatever the peculiar characteristics of the nylB gene sequence, it must also encode a functional, stably folded enzyme, which is another constraint.

Why am I going on and on about nylonase? It has to do with problem of the origin of novelty. Are frameshifts a possible source of new functional information? Might a sequence with alternate frames stay open by chance or be created by chance over evolutionary time? It’s a highly improbable event, but not impossible, I suppose. Might the alternate frames someday be material for frameshifted novel proteins, provided they stay open? They might theoretically be a reservoir for future proteins, but given what we know about the rarity of these kinds of sequences andthe rarity of protein folds in sequence space, the possibility of generating an entire new protein foldfrom a frameshift is extremely, extremely, extremelylow, and would depend on a highly unusual starting sequence tailored in advance for a particular functional specificity. In other words it would need to be designed.

In addition, even should such a sequence exist, it would not long persist in the face of neutral evolution. According to neo-Darwinism there is no magic molecular bouncer who throws out inactivating mutations before they can do their damage to a potential gene. Or to use another metaphor, evolution does not bank potentially useful sequences for future use. For it to do so would require foresight, an idea antithetical to evolutionary theory. Thus, any putative frame-shifted sequences that have been shown to have a functional role are better explained by design than by chance and necessity.

Should anyone disagree with my argument above, I’d like to point out that for a long time it was the standard belief among evolutionary biologists (and geneticists) that random sequence could not generate a functional protein. Frameshifted proteins are almost universally disrupted by stop codons (unless they happen to have an NSF or two like nylB). And even if they aren’t interrupted, the new sequence will be unlikely to fold into a stable protein, given the rarity of functional folds in sequence space (see the first post).

As an aside, as one of the curious facts of history, the disruptive properties of frameshift mutations were used to discover the triplet nature of the genetic code. Says Sir F.H.C. Crick in a lecture on the genetic code he gave in 1964:

This [the ability to combine mutations] has enabled us to tackle the question: is it really a group of three that makes up a codon? The basic idea is the following. We are able to pick up mutants which we believe (from the way they behave in various contexts) are not merely the change of one base into another, but are either the addition or a deletion of a base or bases. What happens when you have a genetic message and you put in an extra base? The reading starts from the beginning until it comes to that point and from there onward the whole of the message is read incorrectly, because it is being read out of phase. In fact we find that these [frameshift] mutants are completely inactive — this is one of our bits of evidence that they are what we say they are. You can pick up a number of such mutants and can put them together, by genetic methods, into the same gene. For example you can put together two of them. Such a gene would be read correctly until it reached the first addition, and then it would be out of phase. When the reading came to the second addition it would [be] read out of phase again, and so the whole of the rest of the message would be read incorrectly. Now it so happens that the left-hand end of this gene is not terribly important for its function. We can actually delete it and the gene will work after a fashion. In this region we have constructed, by genetic methods, a triple mutant, using three mutants all of the same type, and we have found that the gene will nevertheless function fairly normally.

This result is really very striking. Each of the three different faults, used singly, will knock out the gene. You can put them together in pairs in any combination you like, but then the gene is still quite inactive. Put all three in the same gene and the function comes back. We have been able to do this with a number of distinct combinations of three mutants (Crick et al., 1961).

Crick and others found that when three single base frameshift mutations of a particular gene, each completely disruptive on its own, were combined into the same gene, the three insertions together restored the frame enough for the protein to function again! Hence the code must be based on threes.

The sheer improbability of getting a functional enzyme from frameshifted random sequence has been the accepted view for a long time. It is only recently, in the era of big genomic data, that it has begun to be accepted that new proteins dooccasionally arise by frame-shift mutation. The reason? It’s because we find examples in the genome that appear to be products of such events, based on sequence comparisons.

The proteins apparently affected by such frameshifts in the genome are often transcription factors or membrane proteins involved in gene regulation. The apparent frameshift often affects alternative splicing and changes the coding sequence over an exon or so; alternatively, the frameshift affects the end of the protein, resulting in truncation. The fact that such a mutation is located near the protein’s end reduces the amount of disruption to the protein. Many such mutations have been documented to cause disease, however. For a demonstration, just use Google Scholar to search for “frameshift.”

At this point, the chief question that should be in everyone’s mind is, “Can evolution by neo-Darwinian means produce new functional information from frame-shifted sequence? Or are other explanations more likely?”

It boils down to this. Do we say that frameshifted functional proteins are easy to generate, because after all, they exist? Or do we acknowledge that such proteins are not easy to generate and so may be evidence for design?

To reiterate, it used to be standard knowledge that frameshift mutations were always bad. Disruptive. So, for example:

More radical mutational events, such as insertions and deletions that change the reading frame — frameshift mutations — are generally considered to be detrimental (e.g. by causing nonfunctional transcripts and/or proteins, through premature stop codons) and of little evolutionary importance, because they seriously alter the sequence and structure of the protein.

But now it has become popular to offer frameshifts as a quick way to get novelty. I am pretty sure it all began with Ohno, who said:

It has recently occurred to me that the gene started from oligomeric repeats at its certain stage of degeneracy (base sequence diversification) [nylB] can specify a truly unique protein from its alternative open reading frame.

Now the meme has spread. From the Abstract of a paper documenting the “Frequent appearance of novel protein-coding sequences by frameshift translation,” we hear that “Major novelties can potentially be introduced by frameshift mutations and this idea can explain the creation of novel proteins.” And how do they defend the possibility of a functional frameshift? “Some cases of recent evolution of new genes via frameshift have been reported. For example, in bacteria the sudden birth of an enzyme that degrades manmade nylon oligomers was explained by a frameshift translation of a preexisting coding sequence.”

Sigh. The record needs to be corrected. (See my first post, “The Nylonase Story: Where Fact and Imagination Collide.”)

Let us close by considering the nature of the argument being made concerning proposed frameshifts. The fact concerning such proposed frameshifts is that there are sequence similarities between two stretches of DNA, where one part appears to be frameshifted with respect to the other.

Notice that the argument used to explain the appearance of novel genes by frameshift uses a form of inference known as abduction, where one reasons from present effects to past causes.

The surprising fact A is observed.

If B were true, then A would be a matter of course.

Hence, there is reason to suggest that B is true.1

In other words:

The surprising fact of novel genes apparently arising by frameshift is observed.

If it is easy to get new functions from random sequence, then it is a matter of course that frameshifts can produce functional proteins.

Hence it is easy to get new functional proteins from random sequences

Abductive arguments are very weak. The problem is that there can be multiple competing causes that explain the observed effects. The only way to strengthen the argument is to rule out all other competing causes. And design is a particularly strong competing hypothesis. We know design is a cause capable of producing the effect in question, namely the generation of new functional proteins by the addition of frame-shifted code. In fact, given what we know about the rarity of functional proteins in sequence space, as demonstrated experimentally here, here, and here, and theoretically here, design is a betterexplanation than the neo-Darwinian one.

Until someone demonstrates experimentally, in real time, that a frameshift mutation can generate a new functional protein (not just a loss of function) by undirected processes, the inference that it is easy to do so is unjustified. And nylonase is not that demonstration.2

References:

(1) Stephen C. Meyer, Of Clues and Causes: A Methodological Interpretation of Origin of Life Studies. PhD dissertation (Cambridge: Cambridge University, 1990). Charles S. Peirce, “Abduction and Induction,” In The Philosophy of Peirce, edited by J. Buchler (London: Routledge, 1956), 150–154. Charles S. Peirce, Collected Papers, edited by Charles Hartshorne and P. Weiss. 6 vols. (Cambridge, MA: Harvard University Press, 1931–1935).

(2) In a future post, I will discuss experiments that attempt to demonstrate that random sequence can perform simple functions.

Ann Gauger

Editor’s note: Nylon is a modern synthetic product used in the manufacturing, most familiarly, of ladies’ stockings but also a range of other goods, from rope to parachutes to auto tires. Nylonase is a popular evolutionary icon, brandished by theistic evolutionist Dennis Venema among others. In a series of three posts, of which this is the second, Discovery Institute biologist Ann Gauger takes a closer look.

In an article yesterday, “The Nylonase Story: When Imagination and Facts Collide,” I described how some biologists claim that the enzyme nylonase demonstrates that it is easy to get new functional proteins. It has been proposed that nylonase is the result of a frameshift mutation that produced an entirely new coding sequence from an alternate reading frame. I showed why such a claim is false. Now I will explain what that means and something about the unusual properties of the nylB gene that caught molecular geneticist and evolutionary biologist Susumu Ohno’s attention.

What are alternate reading frames? To answer that question, I first need to provide some background information. I will begin by defining some terms I used in yesterday’s post. DNA is composed of two anti-parallel strands of nucleotides. The order of the nucleotides in each strand is what specifies the information the DNA carries. The two strands, called the sense and antisense strands, run in opposite directions. Even though their sequences are complementary, with A always paired with T, and C with G, each strand carries different potential information.

ATG GCA TGC ACC GGC ATT AG → sense TAC CGT ACG TGG CCG TAA TC ← antisense

Before the information in DNA can be used, it must be copied into what we call messenger RNA. The sequence of one strand of DNA, usually the sense strand, is copied using the same base complementarity: G pairs with C, and A with U (U is used in place of T in RNA). We call that copying transcription. The message that has been transcribed from the DNA into that sequence of RNA is now ready to be translated into protein.

Notice the language of information shot throughout these processes. The names for these processes were given by men fully committed to a naturalistic worldview, men such as Francis Crick and Sydney Brenner. Indeed, they were materialists one and all. Yet they saw the parallels between these processes and the human manipulation of text (language) or code (another form of language). The genetic code is the framework that determines the relationship between groups of nucleotides (codons), and the amino acids they specify. The code specifies how to translate the messenger RNA that has been copied or transcribed from the DNA, so that it can be translatedinto a new language, the language of proteins. Below is an illustration of the standard genetic code (source here, used with permission):

Notice that the information in DNA is read in groups of three nucleotides (each group is called a codon), and each codon specifies a particular amino acid. Sometimes more than one codon can specify the same amino acid. For example in the top left corner, the table shows that UUU and UUC both specify the amino acid phenylalanine.

The nature of the code is such that it matters where the first codon begins — the first codon to be read establishes the codon groupings going forward. In the table above the “start” codon is AUG (it also specifies the amino acid methionine). The sequence of codons is “read” by a cellular machine called the ribosome, which starts reading the RNA message at AUG, and then proceeds three nucleotides at a time to translate the message into amino acids. In the sequence below, for example, the first codon to be read would be AUG and that codon determines the frame in which of all the other codons are read.

AUG GCA UGC ACC GGC AUU AGU

Now here’s where it gets interesting. Potentially, DNA can be grouped into different codons, or frames, depending on where the ribosome starts reading. See below for an illustration. For example, the sequence could potentially be read with the groupings shown in frame one (ATG GCA etc.) or frame two (TGG CAT etc., if a proper ATG exists somewhere upstream), leading ultimately to a different amino acid sequences for each. In fact there are six possible ways to group the DNA into codons — three frames on the sense strand going left to right (labeled 1-3), and three frames on the antisense strand (labeled 4-6), going right to left. Below I have laid out the six possible frames for the sequence we began with, but with the alternate frames staggered, and the alternate codons separated by spaces. Notice the sequence stays the same — the only thing that changes from frame to frame is how the nucleotides are grouped. It’s the same sequence, but it could be read and translated differently in each frame. This is because each codon specifies a particular amino acid. Thus, each frame results in a completely different string of amino acids.

frame 1  ATG GCA TGC ACC GGC ATT AG frame 2   TGG CAT GCA CCG GCA TTA G frame 3    GGC ATG CAC CGG CAT TAG

frame 4  TAC CGT ACG TGG CCG TAA TC frame 5   ACC GTA CGT GGC CGT AAT C frame 6    CCG TAC GTG GCC GTA ATC

The codons TAA, TAG, and TGA are stop codons — they specify where the gene ends and protein translation stops. (For extra credit, can you find any ATG or stop codons in the above frames? They are there in both the forward and reverse direction. For more extra credit, can you use the code table to translate different frames, and demonstrate that each frame encodes a different protein?)

So when Venema and others say that nylonase arose by a frameshift mutation that produced a novel protein 392 amino acids long, they are claiming that a completely new coding sequence with frame-shifted codons could generate a functional protein. How likely is that? Not very, given the rarity of functional proteins in sequence space (see my first post). And, as I have already shown in my first post, such an unlikely hypothesis is unnecessary. The nylB gene appears to be the product of a simple gene duplication followed by two stepwise mutations to increase nylonase activity.

There is something special about the nylonase gene’s sequence though, something very odd. nylB has multiple large, overlapping (alternate) open frames that lack stop codons.

How hard is it to get a gene with multiple reading frames?

Let me explain. Roughly one in twenty codons are stop codons. A random DNA sequence will have stop codons about every sixty bases, and may or may not have a start codon. Usually the alternate frames of DNA sequences are interrupted by stop codons. Only the frame that actually specifies the correct gene will have no stop codons at all over a significant length. This system is actually very ingenious. The one frame that needs to be read and translated is identified by an ATG. The other frames will usually lack an ATG and/or will have several stop codons that interrupt their translation, thus preventing the cell from wasting energy on nonsense transcripts.

According to the nylonase story, as told by Ohno and Venema and numerous others, a new ATG start codon was formed by the insertion of a T between an A and G, thus creating a new start codon after the original ATG, which shifted the reading frame for that sequence to that specified by the new ATG, and creating a completely different coding sequence and thus a new protein. Let us grant that scenario for the sake of argument. Normally such a shift would produce a new coding sequence that would be interrupted by stop codons, so the newly frameshifted protein would be truncated. Thus the only reason this frameshift hypothesis for nylonase is even remotely possible is because the sequence coding for nylonase is most unusual, and contains not one, not two, but three open frames Although frameshift mutations are ordinarily considered to be quite disruptive, at least in this case the putative brand new protein sequence would not terminate early due to stop codons.

My point? The first step to getting a new functional protein of any length from a frameshift is to avoid stop codons. The odds of a random coding sequence having an open alternate frame, without stops, are poor. As a consequence, if a protein does have an open frame in addition to its coding sequence, it’s worth paying attention to. And it so happens that nylonase does have more than one open frame. The DNA sequence above illustrates the six frames, numbering them frames 1 through 6. Using that convention, frames 1 and 3 are read from the sense strand. Both have no stop codons over the length of the gene in the sense direction. Frame 4 on the antisense direction has no stop codons either. Frame 1 is the coding frame that specifies the nylonase protein, otherwise known as the open reading frame (ORF). It is defined by the presence of both a start and stop codon. The other two frames have no start codons or stop codons, so I’ll call them non-stop frames (NSFs). They are frames 3 and 4.

The probability of a DNA sequence with an ORF on the sense strand and 2 NSFs is very small. Just exactly howsmall are the chances of avoiding a stop codon in three out of six frames? We set out to determine that by performing a numerical simulation using pseudorandom numbers to generate sequences at various levels of GC content. (By we I mean that my husband, Patrick Achey, who is an actuary, did the programming work, while I determined the parameters.) We chose to vary the GC content because sequences with a higher GC content have fewer stop codons. Remember, a stop codon always has an A and a T (TAA, TAG, and TGA are the stop codons) so having a sequence with a lower percentage of AT content will reduce the frequency of stop codons. Conversely, higher GC content makes the chances of avoiding stop codons and getting longer ORFs much greater, thus also increasing the chances of NSFs. The genomes of bacteria vary in their GC content, from less than 20 percent to as much as 75 percent, though the reason why is not known. One species of Flavobacterium has a genome with about 32 percent GC and 2400 genes — the precise values varies with the strain. The plasmid on which nylB resides is very different. It has 65 percent GC content. The gene encoding nylonase has an even higher 70 percent GC content, which is near the observed bacterial maximum of 75 percent.

We chose to use a target ORF size of 900 nucleotides (or 300 amino acids) because it is an average size for a functional protein. Nylonase is 392 amino acids long; the small domain of beta lactamase, the enzyme my colleague Doug Axe studied, is about 150 amino acids long. The median length for an E. coli protein is 278 amino acids; for humans, the median length is 375.

As expected, the simulation showed that the higher the GC content, the greater the likelihood that ORFs that are 900+ nucleotides long exist. At 50 percent GC, the average ORF length we obtained was about 60 nucleotides; most ORFs terminate well before 900 nucleotides. Indeed, in our simulation only two out of a million random sequences made it to 900 nucleotides before encountering a stop codon. As a result, we could not determine the rarity of NSFs at 50 percent GC — we would probably have to run the simulation for more than a billion trials to get any significant number of NSFs at all.

Sequences at 60 percent GC gave 57 ORFs at least 900 nucleotides long out of a million trials, while sequences at 65 percent GC produced 404 out of a million, one of which also had an NSF.

NSFs were much more probable for sequences that were 70 percent GC, like nylB. In our simulation 3,021 out of a million trials were ORFs at least 900 nucleotides long. That’s a frequency of .3 percent. Of those 3,021 ORFs, 86 had 1 NSF, and none had 2 NSFs. We had to run 10 million trials at 70 percent GC to see any ORFs with 2 NSFs. From those 10 million randomly generated sequences, we obtained 28,603 ORFs; 903 had 1 NSF and only 9 had 2 NSFs.

Interestingly, at 80 percent GC we got a few sequences with 4 NSFs; but I don’t know of any bacterium with a GC content that high.

Our simulation shows that multiple NSFs are very rare. The probability that an ORF 900 nucleotides long with 70 percent GC content will have two NSFs is 9 out of 28,603, or 0.0003. If these figures are recast to include the total number of trials required to get an ORF of that length and GC content and with 2 NSFs, the probability would be 9 out of 10,000,000 trials.

A sequence like nylB is very rare. In fact, I suspect that for all cases where overlapping genes exist, in other words where alternate frames from the same sequence have the potential to code for different proteins, unusual sequence will necessarily be found. Likely it will be high in GC content. Could such rare sequences be accidental? I think that if we compare the expected number of alternate or overlapping NSFs per ORF, with the actual number we will find that there are more of these alternate open reading frames than would be predicted by chance.

From another study of overlapping genes:

Thus, bacterial genomes contain a larger number of long shadow ORFs [ORFs on alternate frames] than expected based on statistical analysis. Random mutational drift would have eliminated the signal long ago, if no selection pressures were stabilizing shadow ORFs. Deviations between the statistical model and bacterial genomes directly call for a functional explanation, since selection is the only force known to stabilize the depletion of stop codons. Most shadow genes have escaped discovery, as they are dismissed as false positives in most genome annotation programs. This is in sharp contrast to many embedded overlapping genes that have been discovered in bacteriophages. Since phages reside in a long term evolutionary equilibrium with the bacterial host genome, we suggest that overlooked shadow genes also exist in bacterial genomes.

Indeed, a study of the pOAD2 plasmid from which nylB came indicates that there are potentially many overlapping genes on that plasmid. nylB′, for example, a homologous gene on the same plasmid that differs by 47 amino acids from nylB, also has 2 NSFs. These unusual and unexpected features of DNA have consequences for how we think about the origin of information in DNA sequences, as I shall discuss in the next post.

Monday: “The Nylonase Story: The Information Enigma.”

by Ann Gauger 

"As is often the case in biology," the scientist and artist David Goodsell has observed, "things get even more complicated when you look closer." And that is the truth.

Yesterday started out as an ordinary Tuesday. Then I set out to read a recent paper published in the journal Cell, "Nuclear Localization of Mitochondrial TCA Cycle Enzymes as a Critical Step in Mammalian Zygotic Genome Activation," by R Nagaraj et al. It reported something rather odd that caught my eye. Very early embryos (at the two- or four-cell stage in mouse or human respectively) undergo a critical transition: they have to go from relying on RNAs and proteins loaded into the egg before fertilization by the mother, to making their own RNA and protein.

The phenomenon is called embryonic genome activation. In order to activate their genomes, embryos have to remove maternal and paternal epigenetic modifications and create new ones appropriate to the embryonic genome.

Stop and think. That's remarkable -- there is a fair amount of information being imparted to the genome by these epigenetic changes, and we know little about how that happens. We do know according, to the authors, that

[s]uch major reprogramming of the genome requires metabolites such as α-ketoglutarate, essential for protein and DNA demethylation, acetyl-CoA required for protein acetylation, [and] ATP for phosphorylation of substrates.

Normally these metabolites are made by specialized enzymes that are part of the tri-carboxylic acid (TCA) cycle. The TCA cycle takes place in the mitochondria, specialized organelles that produce energy for the cell. The mitochondria take up a compound called pyruvate, which is then converted to acetyl-CoA by the enzyme pyruvate dehydrogenase (PDH), and the resulting acetyl-CoA enters the TCA cycle, to produce the other metabolites and ATP.

However, these early stage embryos are metabolically inactive relative to later stages. Their mitochondria are condensed and little enzyme activity is present. So where are the metabolites coming from? Pyruvate is absolutely required for development to proceed. It can come from the fluid in the oviduct -- it can be imported by the embryo. PDH is also absolutely required for development. But where is PDH enzyme activity if not in the mitochondria? And what about the TCA cycle enzymes?

Here's what caught my attention. The enzymes are localized in the nucleus at this early stage of development, where the metabolites are needed for reprogramming the genome, rather than the mitochondria. Now that would be surprise enough -- enzymes that every undergrad biology student learns are in the mitochondria are found in the nucleus early in development. That's remarkable. And now they're involved in embryonic genome activation. But then I dug deeper.

It occurred to me to wonder how hard it would be to get those enzymes into the nucleus, so I looked up pyruvate dehydrogenase and found to my astonishment that it is not one enzyme but an enormous complex of three different enzymatic activities clustered together on a cube-shaped core of 24 units, or alternatively a dodecahedral core of 60 units. The enzymes work together to turn pyruvate into acetyl CoA in a three-step process, handing off to each other as the reaction proceeds. Look here to see pyruvate dehydrogenase described by David Goodsell.

Let me emphasize: this is a core enzymatic activity. The TCA cycle is important to the process by which cells make ATP, the energy currency of the cell. PDH is the link that connects glycolysis, the breakdown of sugars, to the TCA cycle. Without it cells would obtain much less energy from the breakdown of sugars. But it is also essential for embryonic development past the two- to four-cell stage (in mice and humans, and presumably other mammals).

It's also essential for bacteria like E. coli, where it has a similar structure and three-step reaction. This is an ancient enzyme complex, yet of great sophistication.

How could early cells have assembled such a structure, bringing together separate enzyme activities to work cooperatively? Getting enzymes to assemble into multi-subunit structures is non-trivial, requiring multiple side-chain interactions and three-dimensional fit. Even further, the genes encoding these enzymatic activities of the PDH complex are clustered together into a single operon in E. coli. They are neighbors, side-by-side in E. coli's genome, and co-expressed. Of course, that's how an intelligent designer would do it. What's the use of part of a complex? Make the enzymes together and assemble them into a factory to turn pyruvate into acetyl CoA -- it's much more efficient.

I started the day wondering about epigenetics, and uncovered some remarkable things -- reprogramming genomes, apparently mobile enzyme complexes, and operons. Information, complexity and order. It was a not-so-ordinary Tuesday.

Update: Since I wrote this piece I discovered that according to Voet and Voet's biochemistry textbook, PDH complex carries out five enzymatic activities to produce acetyl-CoA, not three, and they state that the PDH complex is the largest eukaryotic enzyme complex known.

By Ann Gauger

The marine iguana (Amblyrhynchus_cristatus), Galápagos Islands, Santa_Cruz

Evolutionists often challenge us for referring to Darwinian evolution as "random." They point to the fact that natural selection, the force that supposedly drives the train, always selects more "fit" organisms, and so is not random. That is only part of the story, though, and to understand why evolution can indeed be called random, the rest needs to be told.

Evolution can be considered to be composed of four parts. The first part, the grist for the mill, is the process by which mutations are generated. Generally this is thought to be a random process, with some qualifications. Single base changes occur more or less randomly, but there is some skewing as to which bases are substituted for which. Other kinds of mutations, like deletions or rearrangements or recombinations (where DNA is exchanged between chromosomes), often occur in hotspots, but not always. The net effect is that mutations occur without regard for what the organism requires, but higgledy-piggledy. In that sense mutation is random

The next part, random drift, is like a roll of the dice that decides which changes are preserved and which are lost. As the name implies, this process is also random, the result of accidental events, and without regard for the benefit of the organism. Most mutations get lost in the mix, especially when newly emerging, just because their host organisms fail to reproduce, or die from causes unrelated to genetics. It can also happen that new mutations are combined with other mutations that are harmful, and so get eliminated.

The random effects of drift are large enough to overwhelm natural selection in organisms with small breeding populations, less than a million, say. New mutations are not born fast enough to escape loss due to drift. There is a fractional threshold in the population that must be crossed before a new mutation can become "fixed," that is, universally present in every individual. A new mutation generally is lost to drift before that population threshold is crossed.

The third part, natural selection, is not random. It acts to preserve beneficial change and eliminate harmful ones. It can be said to be directional. But there are several caveats. Beneficial mutations are rare, and usually only weakly beneficial, so the effects of natural selection are not usually all that strong. Most changes provide only a slight advantage.

In addition, it can happen, and often does, that a "beneficial" mutation involves breaking something, meaning a loss of information, and a loss of potential improvement. This breaking can be irreversible for all intents and purposes. The premiere example in human evolution is that of sickle cell disease. Sickle cell disease is caused by a mutation to the hemoglobin gene that makes red blood cells resistant to the malarial parasite. In one copy the broken gene is beneficial (it increases resistance to malaria), but when two copies are present (both chromosomes carry the mutation), the red blood cells are deformed and cause painful debilitation. The broken gene is actually functionally worse than its normal version, except where malaria is present.

This brings out an important point. Natural selection does not always select the same mutations. The environment determines which mutations are favored. For example, natural selection acts to favor individuals carrying one copy of the sickle cell trait where malaria is present, but acts against the sickle cell gene where malaria is absent. So in this context, selection meanders over a fluctuating landscape of varying criteria for what is beneficial and what is not. Now it is beneficial to carry the sickle cell trait, now it is not. Different populations get favored at different times. In this sense one might say selection has a random component too, because only rarely is selection strong and unidirectional, always favoring the same mutation.

We see this variation in selection with another example, the evolution of finch beaks on the Galápagos Islands. In drought, large beaks are favored, in wet years, small beaks. The weather fluctuates, and so do the beak sizes.

Subpopulations may acquire traits, but because of environmental variation the traits do not become universal. For example, lactose intolerance -- we do not all carry the version of the gene that allows us to digest lactose as adults. Unless suddenly everyone in the world has to eat cheese as a major part of their diet, lactose intolerance won't disappear from our population.

There is a special way evolution can occur -- a sudden bottleneck in the population will tend to fix the traits that predominate in that population. Suppose a nuclear holocaust wiped out everyone except Swedes. The lactose-digesting gene would almost certainly become fixed, as would blond hair, blue eyes, and other Scandinavian traits, provided they ate cheese and lived at high latitudes. Until new mutations in new environments occurred, that would remain the case.

Now you know more about the population genetics of evolution than you imagined could be true. The sum of all these factors is what is responsible for evolution, or change over time. Mutation, drift, selection, and environmental change all play a role. Three out of these four forces are random, without regard for the needs of the organism. Even selection can be random in its direction, depending on the environment.

So tell me. Is evolution random? Most of the processes at work definitely are. Certainly evolution won't make steady progress in one direction without some other factor at work. What that factor might be remains to be seen. I personally do not think a material explanation will be found, because any process to guide evolution in a purposeful way will require a purposeful designer to create it.

One of the most difficult problems for evolutionary biologists to explain is how adaptations manage to appear at the right time for the next stage of evolution to take place. This problem was succinctly summarized by Hugo de Vries, a Dutch botanist. I paraphrase: It's not the survival of the fittest, it's the arrival of the fittest that needs explanation.

To get around the problem, current evolutionary biologists use a word that stands in for this problem: pre-adaptation. In other words, things get ready for what is to come in advance of its arrival. Here's an example. Science Daily reports:

[A] team of scientists from the John Innes Centre, the University of Wisconsin, Madison, and other international collaborators, has discovered how an ancient alga was able to inhabit land, before it went on to evolve into the world's first plant and colonize the earth...

Dr. Delaux said: "At some point 450 million years ago, alga from the earth's waters splashed up on to barren land. Somehow it survived and took root, a watershed moment that kick-started the evolution of life on earth. Our discovery shows for the first time that the alga already knew how to survive on land while it was still in the water. Without the development of this pre-adapted capability in alga, the earth could be a very different place today. [Emphasis added.]

Imagine this scenario: Before plants have colonized the land 450 million years ago, the only stuff on land is an ancient kind of fungus. An ancient algal species, the precursor to modern plants, is living in the sea. For some unknown reason it develops pathways that will enable it to have a symbiotic relationship with the fungus, without which it could not survive on land.

How do we know this? The genes for the symbiotic pathway exist in the modern versions of the ancient alga.

Dr. Delaux and colleagues analyzed DNA and RNA of some of the earliest known land plants and green algae and found evidence that their shared algal ancestor living in the Earth's waters already possessed the set of genes, or symbiotic pathways, it needed to detect and interact with the beneficial AM [arbuscular mycorrhiza] fungi.

With the pathway present, when the time came and the alga splashed up on the shore, it was prepared to share resources with the fungal species that was there. Maybe it had to splash up many times before it had assembled the necessary pathway. Each time, though, it would have died without the symbiotic relationship with the fungus. It was only when the symbiotic pathway was there that the alga could survive on land.

It's difficult to explain how a pathway needed to survive could develop piecemeal, when there is no benefit until the whole thing is assembled. (That's an understatement.) Perhaps it was using the pathway for something else, some would say, and it just happened to provide the means for symbiosis. We just happen to be on the lucky world where the pathway for symbiosis was assembled.

I would tend to discount this pre-adaptationist story (for the original paper go here) because it is so unlikely on Darwinian grounds, were I not an intelligent design proponent. There are numerous other examples that parallel this one, though. We find pre-adapted genes in animals that never have developed the body plans or structures that require those genes. What are they doing there?

When organisms develop the genes and pathways their descendants will use at some future time, before they need them for that particular thing, it's called pre-adaptation, exaptation, or co-option. Whatever you call it, it's either a directed process or an extremely lucky one. Evidence of directed evolution, planning, or any kind of foresight is inimical to Darwinism, so perhaps it's not surprising that the name pre-adaptation is assigned, and it's left at that. But foresight is exactly what we expect from designed processes. Pre-adaptation is another word for preparation -- something all of us should recognize as a hallmark of design.

Breaking news -- another silly article was just published claiming to refute intelligent design, this one at Live Science and drawing on another in Psychophysiology ("Evidence for a vestigial pinna-orienting system in humans"). It's based on the existence of a small muscle behind the ear with no apparent function -- except to tell when we are happy. It can't wiggle the ear, so it is ruled to be vestigial -- left over from previously functional muscles to twitch the ears.

You know the way the argument goes. If we can't find a function, there is none. The only seeming use is as a mood detector. If you are happy the little muscle gets activated, even when you are lying about it.

The argument against design is based on the claim that an intelligent designer wouldn't have done it this way -- he would never create something with no obvious purpose except a residual one. I suppose that's because being all wise and powerful, he wouldn't waste his time with keeping a very small muscle that can't twitch ears. Or so the reasoning goes. The Live Science article quotes the author of the journal article, University of Missouri psychologist Steven Hackley:

"According to intelligent design and creationism, our body was designed by a being with perfect intelligence," he said. "If that were the case, why would he put circuits in our brains that don't work? Why would you put circuits in our brain which are useful for lemurs that are useless for humans?"

This is another case of an evolutionary argument that depends on knowing how an intelligent designer ought to act. I personally don't know the mind of the designer but apparently some biologists (and psychologists) do.

Never mind all the positive evidence for design in the exquisite efficiency of enzyme catalysis, or the amazing sophistication of the human brain, or the optimal sensitivity of the human retina. Ignore those molecular machines, and that information-bearing code in DNA, or the proof-reading and error-correction mechanisms that permit the faithful transmission of information from generation to generation. One small muscle undoes it all.

I don't know. Maybe the designer put it there to make scientists happy -- busy offering silly speculations about how it disproves intelligent design, and then determining what other use can be made of it.

Hold still while I check to see if this makes you happy, and don't wiggle your ears.

 Pictures showing the gradual transition of an ape-like creature to a human one are part of the cultural Gestalt. They are so ubiquitous that people use the idea for jokes. You’ve probably seen the picture describing the evolution of man from a bent-over ape to a bent-over office worker on the computer. But do such pictures represent the truth about our history?

 Let’s consider just one part of the transition that would be required -- to go from an ape-like face to a human one. The portraits of our supposed ancestors are in museums everywhere, and on the web pages of National Geographic -- the latest one, H. naledi, can be found there now. But whether or not that reconstructed face is a part of our history, the question remains: Is there enough time for us to have acquired by purely, Darwinian natural processes [AG1] the differences that make us look human as opposed to ape-like?

 A recent paper purports to tell us how it happened, at least from a genetic perspective. Prescott et al. report the regulatory changes that might have caused the change in morphology (the overall shape) of our faces. On the whole it's an interesting paper with a lot of hard work involved.

 The title, "Enhancer Divergence and cis-Regulatory Evolution in the Human and Chimp Neural Crest," may sound a little cryptic to the uninitiated. Let me provide some background and define some terms. Enhancers are regulatory sequences found in the DNA near the gene(s) they regulate. They act by “enhancing" the expression of those nearby genes. They are also called cis-regulatory elements, hence the "Enhancer Divergence and cis-Regulatory Evolution" part of the title.

 Neural crest cells come from a particular part of the early-developing brain, and they migrate out into the embryo to form the face and other structures in the body. Thus they have a major role to play in embryonic development of the face. Mutations that affect neural crest cells can cause major cranio-facial abnormalities.

 The popular model is that changing enhancers is the way new morphological features evolved, changes such as from an ape-like face to a human one.

 Now for the paper. Prescott et al. began by inducing stem cells of both chimp and human lineages to develop into neural crest cells. (They mostly used in vitro pluripotent stems cells (IPSCs), which are not of embryonic origin. Unfortunately, they did use some human embryo-derived cells to verify that their IPSCs were behaving normally.) From the IPSCs they identified enhancer-like sequences specific to neural crest cells. They then checked for patterns of gene expression driven by those enhancers that differed between chimp and human cells. Having identified those, they then used several methods to examine whether those enhancers did in fact act in a species-specific way in the neural crest. They also checked whether or not the genes they had identified caused human cranio-facial abnormalities when mutated. Some did.

 Essentially, they looked for differences between humans and chimps in their enhancer sequences upstream of genes involved in neural crest development, and found them.

 How convinced am I that their results are valid? They say they identified at least 208 enhancers that work together cooperatively to regulate neural crest gene expression. That I accept, though some of the details may be wrong. 208 is a reasonable number of enhancers to be involved in neural crest specification and behavior. They say that there are probably other enhancers they didn’t identify, and the enhancers they did identify will no doubt have effects in other tissues. Fine. They also report having found multiple enhancer elements upstream of the same gene, each with a positive or negative effect on gene expression that work together. None of this is a surprise -- it confirms the idea that gene regulatory networks are complex and cooperative, and that changing nucleotides in an enhancer (a cis-regulatory element) can change its effect on nearby genes.

 My difficulty is with their claim that anything like this evolved. There is plenty of use of the word evolution in the introduction and discussion sections of this paper. It's actually not surprising that they use evolution so often as an explanatory concept. Evolution is the dominant explanation right now for how things got to be the way they are, so of course these differences in enhancer sequences are explained by evolution…right?

 Let's check the logic. Premise: evolution happens by changing enhancer sequences in a species-specific way. Observation: there are species-specific differences in enhancers. Conclusion: evolution is responsible for these species-specific changes.

 See the problem? That's called circular reasoning. If you accept the premise, then the conclusion is automatic because the two statements are actually saying the same thing. If you don’t accept the premise, of course, you will be unconvinced by this argument. Thus the paper assumes Darwinian evolution without demonstrating that it had anything to do with it.

 So let’s ask the question. Could naturalistic processes guided only by natural selection have accomplished any of this? Changes to the expression of one gene or another can affect morphology -- changing the expression of the gene ALX1  affects the size and shape of finch beaks, for example. Getting one regulatory change to an enhancer is not impossible. But there are multiple enhancers controlling multiple genes involved in facial morphology (see the paper in Nature by Lamichhaney et al.), and to get coordinated regulation of multiple genes, multiple enhancers must have changed in a coordinated fashion.

 Is that possible? Durrett and Schmidt published a paper examining how long it would take to have two coordinated mutations (one inactivating and the other activating) take place in an evolving hominin population. They found it would require in excess of 100 million years. Obviously, the hominin population did not have that long to wait for regulatory change. We supposedly diverged from chimps six million years ago.

 Nowhere in this paper is it demonstrated that anything like this regulatory network could have evolved step by step in the time available. All it establishes is that our faces are different from chimps because we have different regulatory sequences. In fact, I don’t know of anywhere it has been demonstrated that regulatory changes on this scale could be the product of evolution.

 In the meantime, we do know something that can coordinate change on this scale. It’s called intelligence.

(My apologies in advance. None of the italics worked on the web page for Biologic Institute.Assume all genus and species names to have been properly italicized.)

You have probably seen the headlines. Numerous fossils of a new species of hominin have been found in a nearly inaccessible cave in South Africa. Intriguingly, the fossils appear to be part of a very large deposit of bones, apparently left there deliberately by their own kind. Also intriguingly, the skeletons of these individuals may have some traits that are Homo (human-like) and some that are australopithicine (more ape-like). To add further to the mystery, these fossils may be anywhere from 1 to 3 million years old, potentially making them among the oldest fossils identified as being of our genus Homo.

In reading the coverage of Homo naledi, as the species is called now, it seems clear to me that the spin put on the actual bones depends on the assumptions of the writers. What do I mean? Bones can only tell us so much. The rest is a matter of interpretation, and one's point of view inevitably tends to color that interpretation.

Let me give two examples:

The first example is how writers interpret skull size. H. naledi had a small brain compared to ours, about the size of a chimpanzee's. To some writers that seems to indicate the probable lack of high levels of cognition. Only species with brain sizes near our own are considered intelligent. The data used to support that claim are (a) our current knowledge of chimp and gorilla brain sizes, and their lack of rational, abstract thought; and (b) the claim that a gradual progression in brain size exists from australopiths to Homo erectus to Neanderthals to us, indicating a gradual progression in intelligence, which fits the evolutionary story.

There are two problems with these interpretations. First, modern humans exhibit a range in brain sizes, and those differences do not correlate with intelligence. The fossil skulls of Homo erectus, the earliest fossil judged to be "human," exhibit a range in skull size also (but see Casey Luskin's post on Homo habilis). Second, if we were to judge intelligence based on brain size, Neanderthals would be smarter than us since their average brain sizes were more than average human brain sizes. Maybe they were smarter?

H. naledi has a mixture of ape-like and human-like traits. It depends on the desired outcome which traits will be emphasized, and where H. naledi will be placed on the putative fossil tree. The discoverers have placed her squarely on the human side based on her apparent behavior and her inferred ability for long distance walking.

By her apparent behavior I mean the fact that the deposit of bones was found deep in a lightless cave with difficult access. It would take much effort and perhaps some light to make the journey. Yet many, many individuals were found in that dark cave. The discoverers, chief among them Lee Berger, have claimed this as evidence of human intelligence. Such deliberate behavior required significant effort and some danger and indicates a special care for the dead, something that until now only humans were known to exhibit. (I am counting Neanderthals as human.) Yet there are those who seek to claim this is not evidence of ritual behavior or a special care for the dead, that it may be a form of animal behavior. I suspect these people may find the small skull size more significant than the behavior, or they have some other driving motivation for discounting it. Not everyone wants to find evidence of human intelligence and care for the dead in fossils so old or in brain sizes so small, and not everyone likes Lee Berger.

To quote a piece on PBS:

[William] Jungers, [chair of the department of anatomical sciences at Stony Brook University], doesn't dispute that the H. naledi bones belong in the genus Homo and were likely deposited deliberately, but he cautions against "trying to argue for complex social organization and symbolic behaviors." There may be a simple answer. "Dumping conspecifics down a hole may be better than letting them decay around you." He suggests it's possible that there was once another, easier, way to access the chamber where the bones were found. Until scientists can know the approximate age of the Homo naledi fossils, Jungers says they are "more curiosities than game changers. Intentional corpse disposal is a nice sound bite, but more spin than substance." Jungers is more dismissive of Berger's suggestion that we may have inherited the practice of burying our dead from H. naledi, a creature with a much smaller brain than modern humans. "That's crazy speculation -- the suggestion that modern humans learned anything from these pin heads is funny."

So back to the small skulls. The two opposing views must argue that either (a) behavior does not correlate with intelligence because skull size trumps everything, or (b) behavior correlates with intelligence and skull size doesn't matter. In addition, behavior and intelligence trump other differences in morphology. True human status is assigned regardless of shoulder and pelvis shape, or the as yet undetermined age of the fossils.

The second example of spin in operation: Even the way H .naledi is described by science writers reflects a certain bias. The disagreement is not due to scanty evidence but rather to interpretation. Some emphasize that the hands are more chimp-like in their fingers, and that the shoulders appear to be suited for climbing. Thus H. naledi spent time in trees, making her more australopith-like. Others emphasize that limbs and feet appear to be mostly like ours, indicating long distance walking ability, and thus assign her status as Homo.

In fact there is a dispute about whether the find represents one species or two. Berger and his coauthors claim the find represents one species. But others disagree. Jeffrey Schwarz, professor of biological anthropology at the University of Pittsburgh, says this about Homo natali and the subject of classification and preconceived notions:

...How do paleoanthropologists decide if a specimen belongs to a species - whether newly or already named - and if that species is a member of Homo? As the case of Homo habilis illustrates, it is primarily by chronology, not detailed morphology...

Enter the newly announced species, Homo naledi, which is claimed to be our direct ancestor because it has features of australopiths and Homo. Why is it a species of Homo? Because some specimens seem to be like us. Why australopith? Because other specimens have some of their features. Why do all belong to the same species? Because they were found in the same cave. But, the published images tell a different story.... Even at this stage of their being publicized, the "Homo naledi" specimens reflect even greater diversity in the human fossil record than their discoverers will admit.

What to do? As I recently advocated in the journal Science, it's about time paleoanthropologists acknowledged what a taxonomic and undefinable mess the genus Homo has become, and restudy the human fossil record without preconceived notions and the historical weight of overly used names.

In the interests of fairness, as far as I know neither Junger or Schwarz has examined the actual fossils first hand. So to give Berger and his extensive list of coauthors their chance to speak (Berger has welcomed the collaboration of many experts in their fields) I quote from their paper defining the species:

In addition to general morphological homogeneity including cranial shape, distinctive morphological configurations of all the recovered first metacarpals, femora, molars, lower premolars and lower canines, are identical in both surface-collected and excavated specimens ....These include traits not found in any other hominin species yet described. These considerations strongly indicate that this material represents a single species, and not a commingled assemblage....

The collection is morphologically homogeneous in all duplicated elements, except for those anatomical features that normally reflect body size or sex differences in other primate taxa.

This is a very interesting fossil find. I am sure what it means will be argued over extensively in years to come, as more data is collected and analyzed. However, the undisputed bias will undoubtedly be that the evolutionary story of our common ancestry with chimps is true, regardless of where H. naledi is assigned in the story.

Recent research described in a review article in the Journal of Biochemistry & Analytical Biochemistry ("Rethinking the Bacterial Genetic Regulation") indicates there is more to bacterial genetics than was previously suspected.

Bacterial geneticists have had tidy maps of genomes, with promoters and transcription units and operons neatly outlined -- all components that make sure that genes that work together are expressed at the same time. Those things are all still true. But nobody knew until recently about the existence of higher-level dynamics involved in genome regulation.

Using techniques analogous to those of the famous ENCODE project, which opened our eyes to the human genome's complex expression and interaction patterns, scientists have discovered that just like in our genome, much of a bacterium's genome is transcribed in both directions. This opens up the possibility of new forms of gene regulation. In addition, researchers have found the genome is organized into large-scale domains where long stretches of DNA are brought together so that genes involved in related functions can work together.

The way that the DNA supercoils seems to be involved in these dynamic domains. Some genes are preferentially expressed when the DNA is supercoiled, others when it is relaxed. There is also evidence to suggest "rapid movements (snaps) of the chromosomal loci and fast longitudinal density waves fluxing forth and back" along the nucleoid, the region of the cell where the bacterial cell's DNA is located. The DNA is active and moving.

Such dynamic structural change is a remarkable thing.

It may potentially provide global levels of organization that can change in response to the environment in which the cell finds itself. The authors describe one way it may happen in a plant bacterial pathogen, and they suggest that similar modes of environmental response could be involved in other pathogens as well.

Codes within codes within codes -- highly efficient and highly intelligent systems -- don't happen by accident and/or selection. The cell might begin with one code, which is incredible in itself. To layer another code in the opposite direction is far and away beyond that. Then to add a third layer of structural dynamics is simply awe-inspiring.

It has come to my attention that Larry Moran, professor of biochemistry at the University of Toronto, feels left out of the debate about Darwin's Doubt, documented in Debating Darwin’s Doubt. He writes in a post on his blog that he’s “a little miffed” about being omitted from the latter. Moran posted a number of critiques of Stephen Meyer’s book, to which we offered no response at the time. No slight intended, Dr. Moran.

I have now taken a look at the first two of his posts, and noticed some biases. Moran showed a few illustrations that might impress the uninformed, but it's what he didn't say that is revealing.

Trees seeking to represent the history of life are drawn to convey different meanings. Moran's tree shows the diversity of life from a phylogenetic view. Of course, unicellular organisms represent the vast majority of this tree, leaving plants, animals, and fungi to one tiny portion of the tree, indicated by a circle. Moran implies that this indicates the Cambrian explosion was a mere blip. What he doesn’t say is that the Cambrian explosion resulted in the sudden diversity of metazoa, or multicellular organisms like animals and seaweed, for which we have many fossils. As a portion of the diversity on the planet they may be a small fraction, but that doesn't mean the explosion wasn't real.

Take a look at Figure 1 from a paper on the subject ("The Cambrian ‘explosion’: Slow-fuse or megatonnage?") by Simon Conway Morris, an expert in the field. The first fossils for each group are indicated by solid circles. The branch points and putative common ancestors are indicated by asterisks, but have not been observed in the fossil record. To the left are the dates in millions of years. You can read across the tree and see that most of the phyla appear between about 525 and 535 million years ago. That is sudden in geologic time. The few groups present in the Vendian (pre-Cambrian) and still with us today include the fungi, the sponges, and sea anemones. All the rest of the phyla except one, which appears later, first come on the scene in the Cambrian. That is a short time for such basic differences in body plans to appear.

Morris ends his paper with this:

To conclude: The Cambrian explosion is real and its consequences set in motion a sea-change in evolutionary history. Although the pattern of evolution is clearer, the underlying processes still remain surprisingly elusive.

Moran's figure (where apparently a number of phyla have been demoted to classes) reveals the same problem. If you take into account where the circles (the crown groups) are, you will see that the major groups of animals first appear as fossils in the same narrow window. The rest of the branch points are phylogenetic inferences from molecular data, for which no fossils exist.

How did such diverse animals first appear? And how do we reconcile morphological data from fossils (and studies of modern representatives) with the molecular data that keeps pouring in? It is not so easy. A recent paper by Ronald Jenner ("Macroevolution of Animal Body Plans: Is There Science after the Tree?") acknowledges the problem. His abstract:

 A renewed emphasis on the gaps in organization that exist between the crown-group body plans of higher-level animal taxa is a hallmark of the emerging consensus in metazoan phylogenetics. Bridging these gaps is the greatest hurdle that stands in the way of translating our knowledge of phylogeny into a renewed understanding of the macroevolution of animal body plans. Unless a good fossil record is available, there is little hope that we will be able to bridge many of these gaps empirically. We have, therefore, little choice but to resort to our more-or-less informed imagination to produce the historical narratives that are the ultimate goal of our studies of animal evolution. Only by fully engaging with the challenges of devising testable scenarios will we be able to tell where along the spectrum of science and fiction our understanding of animal body plan evolution will finally come to rest. [Emphasis added.]

 Strong language. There's more in the paper:

 The phylogenetic lineages of the major metazoan crown groups are discouragingly short in terms of the number of nodes available along each lineage for resolving the assembly of body plans. This seriously limits the resolution achievable by scenarios, especially in the absence of fossils, but before any evolutionary transformations between body plans can be traced, we must infer hypothetical ancestors. In their recent book on the Cambrian explosion, Erwin and Valentine (2013) likened our attempts to infer the body plans of ancient animal ancestors to séances. The older the nodes in question, the more apt this analogy is. The divergences between most pairs of higher-level crown-group-sister taxa are so significant that the inferences of hypothetical ancestral body plans are generally accompanied by substantial error bars. These become compounded as one integrates the inferences of increasing numbers of hypothetical ancestors to reach deeper and older nodes in the tree. Add to this matrix of uncertainty the potentially limitless play of our imagination, and our attempted explanatory historical narratives may end up being little more than untestable fiction. [Emphasis added.]