Leisola: Cited to Attack Darwin Devolves, Study Devolves on Close Inspection



And continuing with the discussion. Are we building a wall, a fence, a barrier with slats or a bridge to common understanding?



See this note posted on the OP due to our discussion:

Editor’s Note: Because some thought that the Gauger study cited below was being offered as a refutation of the 2012 Näsvall paper (it wasn’t, and it’s not), we have updated the post to make this clear. A much fuller discussion of both papers is now provided here.

Three Misrepresentations on Darwin Devolves' Back Cover
Darwin Devolves: The End of Evolution?
(S. Joshua Swamidass) #2

The ENV article continues the “hit piece” trope. Too bad. This study uses an engineered system to test a specific evolutionary mechanism:

The response over at DI is that this doesn’t count because it is an engineered system. This objection sounds very much like the 3rd law of creationism:

  1. The Law of Reproducible Results : Anything found in nature was Designed, unless it can be reproduced in the lab. Corollary: Anything intentionally done in a lab is not natural; it’s a purposeful result. Therefore, all lab results are evidence of Intelligent Design.

The Ten Laws of Creationism | The Sensuous Curmudgeon

Perhaps @Wayne_Rossiter can explain why he thinks this a sensible response to this study. Moreover, to be clear, this objection does not appear in Behe’s book. From our review, however, we write:

modern evolutionary theory provides a coherent set of processes—mutation, recombination, drift, and selection—that can be observed in the laboratory and modeled mathematically and are consistent with the fossil record and comparative genomics.


Gene duplication followed by divergence is a process that we can observe directly in the laboratory (in systems engineered to test it) and is consistent with the fossil record and comparative genomics (outside the laboratory). It explains how new functional information arises.

Behe's Trainwreck Response to Science
(Neil Rickert) #3

ENV is circling the wagons.

(S. Joshua Swamidass) #4

Yes they are. They are doing it with ad hominems everywhere. It will be interesting in a couple weeks: Gauger and Swamidass: A Conversation About God and Evolution. Could not be more awkwardly timed.

@NLENTS and @cwhenderson, do you have any thoughts on this study?

(Timothy Horton) #5

Interesting to note the DI’s “refutation” of the paper wasn’t published in any scientific journal. It’s a few paragraphs from yet another unreviewed popular press ID nonsense book. :slightly_smiling_face:

They really are living in their own private Idaho.

(S. Joshua Swamidass) #6

I’m very curious how they would justify this:

So, while the described experiments are often promoted as evidence for neo-Darwinian evolution, they either (a) are intelligently designed and do not accurately reflect what happens in nature, or (b) underscore the narrow limits of neo-Darwinian evolutionary change.

What precisely are they claiming does not accurately reflect what happens in nature?

(Blogging Graduate Student) #7

It’s curious that the ENV article decided not to provide a link to the Andersson study directly, instead telling readers which citation in the Science book review article it is. Even worse, they don’t even provide a link to the Science book review article! Almost as though they don’t want their readers to have easy access to the paper, instead having to go out of their way to find it…

In addition to the part about this experiment being “designed”, another criticism of the paper Witt and Leisola make is:

In reality a gene with a weak side-activity was duplicated and the side-activity was strengthened. Intriguing, but nothing more — and nothing new.

They seem to leave out the fact that prior to this “weak side-activity” being duplicated, it had to appear in the first place!
To quote Andersson’s paper (my emphasis):

In a strain lacking trpF, we selected a spontaneous hisA mutant of Salmonella enterica that maintained its original function (HisA) but acquired a low level of TrpF activity sufficient to support slow growth on a medium lacking histidine and tryptophan, representing the innovation of the IAD model (see table S1 for strains). Two mutations were required for this innovation: First, an internal duplication of codons 13 to 15 (dup13-15) gave a weak TrpF activity but led to a complete loss of HisA activity. A subsequent amino acid substitution [Asp10→Gly10 (D10G)] restored some of the original HisA activity ( 10 ). We also isolated two other bifunctional derivatives of hisA that had acquired TrpF activity, but we will not discuss these mutants in this paper (fig. S1, A to C) ( 10 ).

This bifunctional gene was then duplicated many times, amplifying the copy number, and these gene copies then diverged to specialise in one of the 2 functions.

So what the paper actually demonstrated was functional innovation followed by amplification and divergence. Hey, I think I’ve heard that somewhere before…

(S. Joshua Swamidass) #8

See this note posted on the OP due to our thread here:

Editor’s Note: Because some thought that the Gauger study cited below was being offered as a refutation of the 2012 Näsvall paper (it wasn’t, and it’s not), we have updated the post to make this clear. A much fuller discussion of both papers is now provided here.

@art, what do you make of it?

(S. Joshua Swamidass) #9

Can some one from @discovery_institute explain how this is possible? I don’t get it?

(Wayne Rossiter) #10

Frankly, Joshua, you should be ashamed for this additional drive-by, now on me. I didn’t write the ENV piece, nor was I even aware of the paper you cite above. So, why is my name at the front of this thread? It makes it sound like I am aware of any of the contents offered here. I didn’t say anything at all about it. Now, since you seem to be casting a straw man version of me for your audience, let me point you to what I say in my book (as one example):

“For example, the phospholipase A2 (PLA2) proteins secreted in rattlesnake venom are 122 amino acids long (coded by 366 nucleotides), and switching a single nucleotide coding for the amino acid at position forty-nine has a significant impact on their toxicity. That is, the changing of one nucleotide can greatly affect the function of a protein, which itself is relevant to the phenotype of the organism (again, physical form, physiology or behavior). What I have described is the “point mutation” mechanism for adaptive change in organisms.”

It’s been a while back now, but I worked on this system, and published a paper on it:

Gibbs, HL and Rossiter, W. Rapid evolution by positive selection and gene gain and loss: PLA(2) venom genes in closely related Sistrurus rattlesnakes with divergent diets. J Mol Evol. 2008 Feb;66(2):151-66. doi: 10.1007/s00239-008-9067-7. Epub 2008 Feb 6.

So, it’s shameful that you would tar baby me like this. Is this what you call “peaceful”? Is this in “good faith”?

(S. Joshua Swamidass) #11

@Wayne_Rossiter, the article includes this quote:

But biologist Wayne Rossiter emailed with a note about the 2012 Näsvall paper. “For those who weren’t aware, the 2012 study mentioned in the review (reference 11) was… debunked by Matti Leisola (and his co-author Jonathan Witt) in his wonderful book, Heretic: One Scientist’s Journey from Darwin to Design . It’s a must have for anyone’s ID library.”

Are you saying that the ENV article is citing you incorrectly here?

This is not a drive by on you. I’m here to engage. As for using your name, we just found out that that ENV may have inappropriately used your name. Fo this reason, I’ve changed the title now to Leisola, who’s book is quoted.

It isn’t a tarring you, just asking for conversation. I’m sorry it appears that ENV inappropriately used your name. Here are other examples of threads started with someones name:

  1. Ken Keathley: Notes from Dabar and a Baptist's Hope
  2. Clinton Ohlers: Two Parables on Divine Action
  3. Randal Rauser: Is the Atheist My Neighbor?

This is just how we work.Yes, it is peaceful too. It isn’t meant to put you down, but to include you in the conversation. I can see your frustration though, especially if ENV used your name with permission. If this is the case, I hope you can contact them to fix remove your name from the original article.

(Wayne Rossiter) #12

Correction, I did share an email with someone (apparently the author of the ENV post) saying that Leisola had dealt with this 2012 paper.

The point is that this paper is supposed to show gene duplication and adaptation. I never rejected that notion. You used it as an example against Behe, but it’s a rotten one (perhaps my paper would’ve been better!). Leisola is correct. There was no demonstration of “Real-Time Evolution of New Genes by Innovation, Amplification,” only divergence. They artificially made copies of a gene they engineered, and gave it a second (limited) function, so as to cause divergence by selection for function. It’s a terrible example. One I would’ve never used in your Science review.

(S. Joshua Swamidass) #13

Good to know! That is why we ask these questions. So it was appropriate to tag this with your name in the first place. I hope that ENV had permission to include your name in that article.

We do not understand why this rebuttal is valid. Even the editor’s note from ENV seems to disagree with you.

Can you explain to us why this is “rotten”? We can discuss your paper next if you are interested (but I don’t want this to mushroom unless you want it).

(S. Joshua Swamidass) #14

We understand it as an excellent example. It is a system engineered to test a specific evolutionary mechanism. The @AGauger study was misunderstood by Leisola, and she said as much here: https://discourse.peacefulscience.org/t/gauger-realtime-evolution-by-innovation-amplification-and-diversification/4828 . So, when you have time, I hope you can come explain why you think it is not a valid study to cite.

@Art and @Nlents, what do you make of this?

(Wayne Rossiter) #15

Ann didn’t feel that her work demonstrate any refutation of the Nasvall paper. That wasn’t its intent. I didn’t quote that from the book (the ENV author did). The refutation is in the actual methods Nasvall et al. used, and what the paper claims. I’ve explained the problem above. For this reason, it’s a terrible example of gene duplication/adaption/divergence, but you used it again Behe, as if he should’ve dealt with it (among the thousands of other possible papers you could’ve used). It’s simply whack-a-mole tactics. Roll through some rolodex of examples, and then complain that someone hasn’t addressed all of them. Ken Miller is notorious for that crap. Don’t be like that.

(S. Joshua Swamidass) #17

Very helpful clarification. Thank you @Wayne_Rossiter. That is our point, my point it seems. @Agauger did not think her paper refutes Nasvall, but Leisola did. Perhaps you agree with Leisola, and disagree with @Agauger, because you did point the ENV author to this excerpt.

By which I think you mean:

So, do you disagree or agree with Behe’s response to this?

His objection is that it is an engineered system, therefore it is an example of intelligent design. That is also what @Agauger thinks too. Are you agreeing with them in that reasoning or disagreeing? I’m asking to tease out what you really mean here.

From what we meant to explain from that article, it is an excellent citation. I’m not sure you why you think is not a valid citation yet. It is a system engineered to test a specific mechanism of evolution, and finds that this mechanism is very effective. The fact it is engineered, does not some how make its findings about the evolutionary mechanism irrelevant.

(Mikkel R.) #18

Sorry but I don’t agree with this portrayal.

They didn’t “artificially” make copies of it, the bacteria made copies of it. The way you describe it seems to imply the researchers literally made and inserted extra copies of the gene to synthetically mimic the process of duplication. But that’s not what they did. They placed the gene in a location on a plasmid more-prone-than-“normal” to duplication. But that duplication is done in bacteria, by the bacteria, not artificially by the researchers.

You couldn’t say that placing the gene in a genetic location with a higher-than-average rate of duplication for the genome is to “artificially make copies” of it, as all they did would have affected the rate of change, not forced the possibility or the occurrence of change.

But that was only done so as to be able to see these changes in a practical timeframe, as opposed to having to wait years for similar amounts of duplication to occur had the gene been in a less duplication-prone area on a plasmid, or in the bacterial chromosome.

They also didn’t engineer the gene to “give” it a second limited function, they selected it under conditions that would favor those kinds of mutations should they actually occur.

(Arthur Hunt) #19

Following up on @Rumraket’s remarks:

Nasvall et al. developed and set out to test a model for the evolution of new genes that they call the “innovation-amplification-divergence (IAD)” model:

We propose the innovation-amplification-divergence (IAD) model (Fig. 1A), which allows the evolution of new genes to be completed under continuous selection that favors maintenance of the functional duplicate copies and divergence of the extra copy from the parental allele (5). The IAD model proposes that the ancestral gene has a weak secondary activity (innovation) (6, 7), and when a change in conditions makes this activity useful, selection favors increased gene dosage (amplification), resulting in two or more copies of the parent allele. The increased copy number pro- vides multiple targets for beneficial mutations and buffers any negative effects a new mutation may have on the original activity. During continuous growth under conditions that select for both the original activity and the new activity, beneficial mutations will accumulate (divergence) in the copies. Any improved copy can be further amplified, whereas less functional copies, including the parental gene, can be lost. Ultimately, this results in a gene duplication in which one gene copy encodes the parent activity and another copy provides an improved, new activity.

To test this model, they broke it down into its three components and evaluated each step separately. For this, they focused on two enzymes that catalyze chemically-related reactions, but that are in pathways for histidine or tryptophan biosynthesis, respectively.

To experimentally test the IAD model, we examined a histidine biosynthetic enzyme (HisA), and through continuous selection we created, by duplication and divergence, a new gene that catalyzes a step in tryptophan synthesis. The original HisA and TrpF enzymes both catalyze isomerization of a phosphoribosyl compound, but each acts on different substrates in the biosynthesis of the amino acids histidine and tryptophan (Fig. 1B). HisA and TrpF enzyme activities are selectable by growth in minimal media lacking histidine and tryptophan. In addition, the enzymes are structurally related and evolved from a common ancestor (8).

Thus, briefly, they first asked whether they could identify a HisA variant that could catalyze the TrpF reaction; this was simple, accomplished by screening trpF mutants that could grow on media lacking tryptophan (and, obviously, histidine). They isolated bacteria able to do this and confirmed that they had indeed isolated strains in which HisA had acquired TrpF activity (and was thus bifunctional). This is a positive experimental confirmation for the “innovation” part of their model. (Note that, other than conduct the experiment with a mutant strain, these researchers did nothing to promote or guide the outcome.)

The researchers then set out to test the second step of the model, and specifically to ask if gene amplification could improve the growth of cells that carry the bifunctional gene. I won’t describe the set-up, other than to say that they placed the bifunctional gene into a genetic context that allowed for easy generation and assay of gene amplification, and then performed selections in strains in which the bifunctional strain was the only source of HisA and TrpF activity. They identified variants that indeed grew better, and confirmed that the improved growth was due to gene amplification. This outcome is positive experimental support for the second part of their model.

To test the third step of the model, they then allowed lineages to continue to evolve under the same selection (no his or trp) for thousands of generations. They succeeded in isolating many better-growing variants. Different of these had bifunctional enzymes with further alterations; collectively, these alterations define many (most?) of the different mechanisms by which evolution might improve the enzyme:

As predicted from the IAD model, we observed the appearance of a diverged gene copy with improved activity, relaxed selection for maintenance of the unimproved copies in the amplified array, loss of the unimproved copies, and, in some cases, reduction in the total gene copy number.

Compare what Nasvall did with Ann’s description:

The story: Näsvall et al. created a gene that encoded an enzyme that was able to carry out two functions, but very poorly.

No, they did not “create” a bifunctional gene. Random mutation and natural selection was the mechanism, not design.

They placed that gene in a strain of Salmonella that lacked the genes for those functions. Then they cultured the bacteria under conditions where they needed to carry out those functions to grow. Guess what? The bacteria duplicated the genes so as to make more of the poorly functioning enzymes.

Which was the exact point of the assay – to test the hypothesis that duplication can contribute to enzyme evolution. The answer, it turns out, is yes. (It seems to me that Ann’s objection here is quite akin to claiming that enzyme assays are not valid because the biochemist first makes extracts (and purified enzymes) and then conducts tests with non-natural substrates. In other words, the objection fails.)

Over time, the genes acquired mutations (remember, we are talking about a lot of bacteria) and the ones that helped the most gave the best growth rate, and… after thousands of generations they had evolved separate enzymes for each function — meaning not that they had evolved new enzymes, but that they had divided up the pre-existing dual-functional gene into two genes.

Again, not correct. Nasvall et al. noted:

The evolved genes fell into three classes: (i) specialized genes with strongly improved HisA activity and loss of TrpF activity, (ii) specialized genes with strongly improved TrpF activity and loss of HisA activity, and (iii) generalist genes whose encoded enzyme showed a moderate increase in both activities …

Note that, among the classes of enzymes are ones in which HisA has now become TrpF (for all intents and purposes). This is a transition that Gauger and Axe say cannot occur (from the abstract):

Considering that Kbl2 and BioF2 are judged to be close homologs by the usual similarity measures, this result and others like it challenge the conventional practice of inferring from similarity alone that transitions to new functions occurred by Darwinian evolution.

Note that the only “tools” Nasvall et al. used were random mutation and natural selection - exactly what Gauger and Axe claim are not tenable means for generating new functions.

The bottom line - Ann, @Wayne_Rossiter, Behe, and others object to Nasvall et al. for reasons that are completely baseless.

(Sorry this is so long - the words are needed to clear up the confusion sown by Ann and others.)

Gauger: Answering Art Hunt on Real Time Evolution
(Dave Carlson) #20

Don’t apologize. That was a fantastic summary.

(S. Joshua Swamidass) #21

@Wayne_Rossiter messaged me that he plans to respond to @art’s post in the coming week or so. Please keep this thread free of clutter in respect of his willingness to engage. Treat him with respect. There are several questions he has been asked, a lot of text written to him. It may take him a few post to make his point. Even if we disagree, we should try and understand him first.