Is science just another religion? (and so much more..)

Multiple evolution simulators use this technique with interesting results.

http://avida.devosoft.org/

http://www.telegraph.co.uk/news/sci...simulated-life-forms-evolve-intelligence.html


~~ Paul
The trouble is, the devil is in the details. My guess is that the 'computer code' would be some sort of code that is designed to be rich in meaningful primitives - like "turn left", or "stop and taste" or whatever - which is not analogous to real computer code or real DNA. Put a slightly different way, it is easy to sort of build the answer into experiments like this - possibly without even realising what you are doing. That is why I think using real, 32 or 64-bit binary code, would be a lot more realistic.

Looking at this page, I get the impression that the code operates on a restricted set of instructions that may be packed with useful operators.

If you have the time, why not download the AVIDA software and experiment for yourself!

If you get somewhere interesting, please set up a thread here to report and discuss your results!

David
 
The trouble is, the devil is in the details. My guess is that the 'computer code' would be some sort of code that is designed to be rich in meaningful primitives - like "turn left", or "stop and taste" or whatever - which is not analogous to real computer code or real DNA. Put a slightly different way, it is easy to sort of build the answer into experiments like this - possibly without even realising what you are doing. That is why I think using real, 32 or 64-bit binary code, would be a lot more realistic.

Looking at this page, I get the impression that the code operates on a restricted set of instructions that may be packed with useful operators.

If you have the time, why not download the AVIDA software and experiment for yourself!

If you get somewhere interesting, please set up a thread here to report and discuss your results!
I posted one article that discusses some of the results.

Yes, there are probably some useful operators in Avida. So now we have to argue whether those useful operators could arise from less useful ones. But as we work backward toward simpler and simpler primitives, we get to the point where evolution can explore a large part of the sequence space in a reasonable time. As Bart posted:

http://rsif.royalsocietypublishing.org/content/5/25/953

Keep in mind that protein superfamilies are clustered together in sequence space, reducing the distances that evolution has to travel.

So where is the point where evolution reaches the brick wall?

~~ Paul
 
I posted one article that discusses some of the results.
Can you give a link?
Yes, there are probably some useful operators in Avida. So now we have to argue whether those useful operators could arise from less useful ones. But as we work backward toward simpler and simpler primitives, we get to the point where evolution can explore a large part of the sequence space in a reasonable time.
This is really critical - I mean if (say) you have a collection of "useful operators" evolution becomes massively easier, whereas if you have to spell them out in computer code or DNA things get a lot tougher in a combinatorial way.
As Bart posted:

http://rsif.royalsocietypublishing.org/content/5/25/953

Keep in mind that protein superfamilies are clustered together in sequence space, reducing the distances that evolution has to travel.

So where is the point where evolution reaches the brick wall?

~~ Paul
When Bart's link seemed pretty extreme - for example, as I pointed out above amino acids have more relevant characteristics than just hydrophobic/hydrophilic.

Somewhere I have seen some actual experiments with proteins with point mutations - I can't just find the reference. Anyway the point was that a single mutation nearly annihilated the functionality, and that was completed for two changes - in other words at that 'distance' you had entered the desert where NS can't help.

I'd like to check if those mutations involved amino acids that could remotely be considered similar.

As to where evolution hits a brick wall, the problems don't even seem to end with creating the right sequence - multiple proteins may be required to achieve something useful, together with some sort gene regulation! Many newly evolved proteins would surely suffer strongly negative selection pressure because without regulation they would be worse than useless.

I mean, NS was devised to explain evolution assuming that it operated in situations where there weren't these huge combinatorial explosions. I used to accept it on that basis - and used to try to argue for it here on SKEPTIKO!. It was the various conversations we had with LoneShaman that persuaded me that most of the steps in evolution simply can't be guided by natural selection, and so there is a problem - at least in the case of the early stages of life when most of the proteins first came into existence.

It is also worth remembering these articles that I showed you a while back:

http://www.thenewatlantis.com/publications/getting-over-the-code-delusion

This illustrates that there is a lot more complexity in the system around such things as just how the DNA winds round the histones etc!

As far as I can see, all that extra complication only makes evolution by natural selection even less plausible - even though Stephen Talbott isn't arguing that case.

David
 
I really am not qualified to make a lot of sense of the technical details in these papers (not being sarcastic here, David, just honest) but, since Bart's link contains a criticism of Axe I think it only fair to post his (among others) rebuttal.

https://uncommondescent.com/intelli...allan-miller-and-to-dryden-thomson-and-white/
Thanks for that!
Could proteins have developed naturally on Earth, without any intelligent guidance? The late astrophysicist Sir Fred Hoyle (1915-2001) thought not, and one can immediately grasp why, just by looking at the picture above, which .....

At least that quote maybe justifies what I am saying (and what others here and elsewhere argue) - I'm not just shooting my mouth off with a naive argument!

Also from that paper there is a link to this:

https://www.ncbi.nlm.nih.gov/pubmed/10966772

Which is the research I wanted to mention in reply to Paul - work that showed that any changes to the protein sequence are massively damaging. This isn't really surprising - I mean the initial protein molecule folds in a variety of local ways, and then folds in a 3D way to obtain the final molecule. Even tiny changes can spoil that process, and the result just doesn't work.

David
 
Last edited:
Can you give a link?
http://www.telegraph.co.uk/news/sci...simulated-life-forms-evolve-intelligence.html

Somewhere I have seen some actual experiments with proteins with point mutations - I can't just find the reference. Anyway the point was that a single mutation nearly annihilated the functionality, and that was completed for two changes - in other words at that 'distance' you had entered the desert where NS can't help.
I don't believe that point mutations typically eliminate the function. Can you find the reference?

As to where evolution hits a brick wall, the problems don't even seem to end with creating the right sequence - multiple proteins may be required to achieve something useful, together with some sort gene regulation! Many newly evolved proteins would surely suffer strongly negative selection pressure because without regulation they would be worse than useless.
But why do you think it was this complicated early on?

~~ Paul
 
http://www.telegraph.co.uk/news/sci...simulated-life-forms-evolve-intelligence.html


I don't believe that point mutations typically eliminate the function. Can you find the reference?
I just did!
https://www.ncbi.nlm.nih.gov/pubmed/10966772

The problem is in the way proteins are created. They start as a linear sequence, then they fold in local ways - helices and so forth, and then they fold again to give them their vital tertiary structure. Every bit of the protein is presumably relevant to that tertiary folding process.

Sorry this is behind a paywall to which I don't have access, but I think it relates to work that I have seen . This is probably a better way into this issue.
But why do you think it was this complicated early on?

Well of course, really early on it was incredibly complicated because you had to get enough machinery going to sustain life before there was life or anything to survive if it was fit! But I mean even a worm presumably has most of that complexity - indeed the research to which he refers was probably done on such simple organisms.

I guess what exasperates me, is that the whole story of evolution by natural selection is presented to the public as a simple proven fact - slam dunk, end of story. The truth is so far from that, and yet the public have to be fed this story without any trace of doubt creeping into the discussion, and those who express such doubt have to be crushed as effectively as possible. What is wrong with saying "we don't yet know if NS could achieve this", or even "there are doubts as to whether all this can be achieved by natural selection".

As regards the Telegraph link, even though it is pretty vague, it illustrates exactly what I was discussing before. Those 'neurons' will (at least I presume) not get specified indirectly by something remotely like DNA, they will be specified by parameters that directly relate to their layout and structure - which makes the combinatorial tasks of searching by NS, vastly easier. I think the best equivalent of doing evolution in the computer is to start with something like binary computer code - something whose structure is only indirectly related to the task it will perform.

Furthermore, I reckon you have been around in the software industry roughly as long as I have, and you have probably read hundreds of reports like this. They are described in terms of promising preliminary results, but then you never really hear of them again.

David
 
Last edited:
Last edited by a moderator:
Sorry Paul,

I have done it again - edited your post when I was trying to reply to it!

Anyway, I will try to get a clarification from Douglas Axe.

The problem is to find his email address.

Edit I have now sent him a query - so I will let you know.

David
 
Last edited:
Paul,

Douglas Axe has sent a detailed reply, that rather confirms what you say, but puts a different interpretation on it - let's discuss it:

Hi David—
You may be mixing up multiple experiments. I have described single amino-acid substitutions that eliminate the function of barnase (Biochem 37(20):7157-7166, 1998). But these are rare substitutions, amounting to only 5% of the substitutions explored.
I don’t have time to read the Panda’s Thumb post you refer to, but the JMB paper in question (JMB 310:585-595, 2000) shows complete inactivation of barnase when 15 of 84 exterior positions are substituted with highly conservative alternatives, and dramatic impairment of beta lactamase when 30 of 171 exterior positions are substituted in the same way.
In assessing the implications for evolution, you have to consider how much sequence information must be present for function, not how many positions in a wild-type enzyme can be altered. The analogy to written language (as used in the paper) illustrates this well.
Fir eksemple, wif sum werk yu kan figyer owt thuh meening uv thiz centinse eevn thowe it haz lotz uv tiepoze.
If the typos in that sentence were completely random it would be unreadable, but by preserving phonetics (analogous to preserving amino-acid properties in the JMB study) the changes leave it readable. Still, while the above sentence remains readable despite having more than 30 typos, this doesn’t at all make it possible to convey the meaning by blind typing. Blind evolution can’t produce anything like that sentence.
I hope this is of help, and yes, feel free to quote.
Regards,
Doug
David
 
Again, this insistence that a complex protein cannot poof into existence is pointless. Everyone agrees. The question is, how did the protein evolve from simpler proteins? And it didn't have to do it step by step. Perhaps domains of the protein were coded by separate genes and then the genes came together by transposition. Also, keep in mind that many of these proteins or sub-proteins could have evolved separately, allowing for visiting more of coding space. Then they could come together by horizontal transfer.

Here's an interesting article:

http://www.pnas.org/content/112/52/15976.full

~~ Paul
 
Again, this insistence that a complex protein cannot poof into existence is pointless. Everyone agrees. The question is, how did the protein evolve from simpler proteins? And it didn't have to do it step by step. Perhaps domains of the protein were coded by separate genes and then the genes came together by transposition. Also, keep in mind that many of these proteins or sub-proteins could have evolved separately, allowing for visiting more of coding space. Then they could come together by horizontal transfer.

Here's an interesting article:

http://www.pnas.org/content/112/52/15976.full

~~ Paul

Well as you can see from the above letter, Ax's experiment didn't seem as significant as I had thought!

However, the crucial point which we can presumably agree on, is that back when Darwin's theory was invented, NS was not supposed to have to alternate with chunks of evolution that had to proceed combinatorially?

The obvious question is whether NS can reasonably work when it has to alternate with combinatorial evolution (CE). Alternatively, are you suggesting that there are long chains of proteins in 'protein space' and that evolution can just hop from one to the next without getting mired down in CE?

In the past, when I supported evolution by NS (without giving the matter that much thought) I hadn't realised that there were chunks of evolution that have to rely on CE. I thought that arguments based on CE were made by those who hadn't understood how NS is supposed to work!

David
 
Last edited:
However, the crucial point which we can presumably agree on, is that back when Darwin's theory was invented, NS was not supposed to have to alternate with chunks of evolution that had to proceed combinatorially?
I'm not sure what natural selection has to do with the random (or possibly not so random) walk that a gene takes through coding space. The latter is a search mechanism while the former is a filter mechanism.

The obvious question is whether NS can reasonably work when it has to alternate with combinatorial evolution (CE). Alternatively, are you suggesting that there are long chains of proteins in 'protein space' and that evolution can just hop from one to the next without getting mired down in CE?
Natural selection is happening all the time. A random (?) change occurs in a gene and then it (a) doesn't really affect the organism; (b) is awful and gets stomped; or (c) is beneficial, possibly only slightly. And there is nothing stopping those low-affect mutations from fixing in the population.

In the past, when I supported evolution by NS (without giving the matter that much thought) I hadn't realised that there were chunks of evolution that have to rely on CE. I thought that arguments based on CE were made by those who hadn't understood how NS is supposed to work!
But that would suggest that natural selection is selecting which mutations occur. Is that what you thought?

~~ Paul
 
I'm not sure what natural selection has to do with the random (or possibly not so random) walk that a gene takes through coding space. The latter is a search mechanism while the former is a filter mechanism.
Well look, a standard idea in evolution is that the gene for a protein gets copied (by chance) and then the copy undergoes mutations guided by natural selection that ends up producing another protein that is good for something else. The real question here is that on the way between P1 and P2, how many of the steps will be of type (a) for the organism - i.e. they will make no difference because the gene neither codes for a useful P1 nor does it code for a useful protein P2. I am curious, are you saying that the whole transition can be achieved using steps that are guided by NS, or do you accept that at least some of the steps will be unguided?

I am trying to figure out just how much common ground we have here!

David
 
Well look, a standard idea in evolution is that the gene for a protein gets copied (by chance) and then the copy undergoes mutations guided by natural selection that ends up producing another protein that is good for something else. The real question here is that on the way between P1 and P2, how many of the steps will be of type (a) for the organism - i.e. they will make no difference because the gene neither codes for a useful P1 nor does it code for a useful protein P2. I am curious, are you saying that the whole transition can be achieved using steps that are guided by NS, or do you accept that at least some of the steps will be unguided?
Every step is guided in the sense that a highly deleterious change will cause the organism to die.

Ignoring that, I suspect that most changes are basically neutral and so the genes change by genetic drift, with no real guidance by selection. Occasionally a helpful change occurs and selection gets a chance to capture it, though that is, of course, not guaranteed.

If P1 and P2 can evolve more or less independently, then P1 can maintain its original function while P2 has a chance to evolve a new function.

We have to be careful not to think of natural selection as a force. It's just a result of mutation and differential reproduction in organisms with heredity.

~~ Paul
 
This is strange, you simply dismiss a (real) peer reviewed paper in a respectable journal, and reply with a blog post from a blogger we know nothing about.
Which is going to be more out on a limb?

But if we look at what Boammaaruri does, it is exactly the same as what the DI does, it takes Axe's 2004 paper “Estimating the Prevalence of Protein Sequences Adopting Functional Enzyme Folds”, and it takes the 1 in 10(77) figure from that paper, then that figure is extrapolated to every protein in existence, and voila evolution is not possible.
Do you really think we can do that?

Well to find out, let us concentrate on Axe's 2004 paper, which is an actual peer reviewed one, unlike the ones published in BIO-complexity, let's take a look at what it actually says.
Does this paper allow for such a broad generalization?

Well according to this essay, it does not.
It summarizes:

Summary

To summarize, the claims that have been and will be made by ID proponents regarding protein evolution are not supported by Axe’s work. As I show, it is not appropriate to use the numbers Axe obtains to make inferences about the evolution of proteins and enzymes. Thus, this study does not support the conclusion that functional sequences are extremely isolated in sequence space, or that the evolution of new protein function is an impossibility that is beyond the capacity of random mutation and natural selection.

So, the whole “Hard-to-Get-a-Protein” hypothesis (HGP for short, a term i borrow from this essay: Is it easy to get a new protein?) rest on the assumption that a Axe's estimate for one very specific protein, under very specific, and specially selected, experimental circumstances, is valid for the evolution of any new protein.

Further, given that the HGP hypothesis is a negative argument (it is impossible that...), it is extremely sensitive for any positive arguments for the opposite.
And if we look for scientific evidence for that, it is not that hard to find.

For instance the paper that you dismissed one one aspect:
In your link, they got a really low estimate of 10^10 by assuming that amino acids could be coarsely grouped into hydrophilic ones and hydrophobic ones! Well, I am sure you could reduce the number of amino acids slightly (if starting life all over again), but those amino acids vary in a whole variety of ways - for example their bulk, and whether they contain SH groups etc. The shape of the amino acids is critical because the protein has to fold twice, to become active (there may be exceptions, I don't know).

David
While i have to admit that a lot of these papers go partially over my head, the coarsely grouping amino acids into hydrophilic and hydrophobic ones, is only one suggestion to reduce sequence space, if i understand correctly.

For instance the suggestion that folds with fewer amino acids form scaffolds capable of supporting all protein functions :
The assumption that a protein chain needs to be at least 100 amino acids in length also rather inflates the size of sequence space when it is known that many proteins are modular and contain domains of as few as approximately 50 amino acids thereby reducing the space to 2050or approximately 1065 (e.g. Sobolevsky & Trifonov 2006). The conclusion from all of these coarse-graining approaches is that a reduced alphabet of amino acids is quite capable of producing all protein folds (approx. a few thousand discrete folds; Denton 2008) and providing a scaffold capable of supporting all protein functions

Or:
that the actual identity of most of the amino acids in a protein is irrelevant. An example in nature could be the prokaryotic DNA methyltransferases which each contain a target recognition domain (TRD) of approximately 150 amino acids that recognizes specific DNA sequences usually of 3–6 bp in length, and a conserved catalytic domain. The thousands of known TRD sequences show negligible amino acid sequence conservation despite the rather limited number of nucleotide sequences they are required to recognize.

More evidence for a more navigable sequence space:
Functional proteins from a random-sequence library
In conclusion, we suggest that functional proteins are suf®ciently common in protein sequence space (roughly 1 in 1011) that they may be discovered by entirely stochastic means, such as presumably operated when proteins were ®rst used by living organisms. However, this frequency is still low enough to emphasize the magnitude of the problem faced by those attempting de novo protein design.

Simple evolutionary pathways to complex proteins

In summary, the conclusions derived from the current study are based on a model that is quite restrictive with respect to the requirements for the establishment of new protein functions, and this very likely has led to order-of-magnitude underestimates of the rate of origin of new gene functions following duplication. Yet, the probabilities of neofunctionalization reported here are already much greater than those suggested by Behe and Snoke. Thus, it is clear that conventional population-genetic principles embedded within a Darwinian framework of descent with modification are fully adequate to explain the origin of complex protein functions.

Another interesting read:

Exploring protein fitness landscapes by directed evolution

Despite the vast size of sequence space and the complex nature of protein function, the Darwinian algorithm of mutation and selection provides a powerful method to generate proteins with altered functions. This simple uphill walk on a fitness landscape in sequence space works because proteins are wonderfully evolvable and can adapt to new conditions or even take on new functions with only a few mutations.

And even more positive evidence against the HGP hypothesis comes from examples of evolution in action.
As there are the appearance of Nylonase, and the E. coli long-term evolution experiment

So to conclude, in my opinion, the HGP hypothesis being a negative argument, can only work if it would have to deal with known facts. To prove a negative it would have to be only vulnerable to unknown unknown's.

For instance, an object with mass can not hover against gravity, unaided by an external energy source pushing against it. To claim that statement wrong we would have to show new information, information that we now do not even know we lack.

The HGP hypothesis is not only vulnerable to unknown unknowns, it is also vulnerable to known unknowns, and to known facts.

As i think is shown, with the positive evidence for a smaller sequence space with easier pathways to functional proteins, the HGP hypothesis fails.

I would agree that the estimates for size of protein space, and the chance of getting a functional protein, are completely open for (reasonable) discussion.

But that is not the argument that the DI is making, they stick to their 1 in 10(77) figure, which is at the most extreme, in the light of the other evidence, the implausible end of the spectrum.

So i ask you, David, do you think the DI will ever let that figure go, will "evolution news" ever publish a blog post that says "in light of recent scientific evidence, we were wrong on that 1 in 10(77) thing"?

Edit: another good read:

How do new proteins arise?

Conclusions
We have reviewed here recent insights on how new
proteins emerge. We suggest the number of newly arising
proteins, for example from intergenic regions, is very
small and unlikely to be a major driving force in organismic
evolution. The underlying reason may be that
forming functional proteins or transforming structures
into each other is very difficult, albeit not impossible.
Latent traits, multistable bridge structures and disordered
regions seem to play a major role in transitions and
certainly deserve further attention for specific experiments.
Rearrangements of domains are frequent and
can be seen as a way to circumvent the tight biophysical
constraints imposed on protein evolution. Current
research indicates that rearrangements are frequent and
follow a random model [8]. Thus rearrangements can be
seen as a ‘higher level’ form of neutral evolution: it allows
for the exploration of ‘rearrangement space’, which is
reminiscent of the concept of neutral exploration of
sequence space by point mutations, until rare beneficial
‘mutations’ (arrangements) become fixed.
 
Last edited:
And even more positive evidence against the HGP hypothesis comes from examples of evolution in action.
As there are the appearance of Nylonase, and the E. coli long-term evolution experiment
It turns out that the Nylonase story is a bit more complicated!


https://www.evolutionnews.org/2017/05/the-nylonase-story-when-imagination-and-facts-collide/
Venema is right. If the nylonase enzyme did evolve from a frameshifted protein, it would genuinely be a demonstration that new proteins are easy to evolve. It would be proof positive that intelligent design advocates are wrong, that it’s not hard to get a new protein from random sequence. But the story bears reexamining. Is the new protein really the product of a frameshift, or did it pre-exist the introduction of nylon into the environment? What exactly do we know about this enzyme? Does the evidence substantiate the claims of Venema and others, or does it lead to other conclusions?

If the nylonase story as told above were true, namely that a frameshift mutation resulted in the de novo generation of a new protein fold with a new function, it would indeed constitute a substantial refutation to Meyer and Axe’s claim. If a frame-shift mutation can produce a random new open reading frame in real, observable time, and give rise to a new functional enzyme, then it must not be that hard to make new functional protein folds. In other words, functional protein folds must not be rare in sequence space. And therefore Stephen Meyer’s arguments about the difficulty of getting enough new biological information to generate a new fold must be wrong as well. Venema flatly asserts:

Just to be clear, a frameshift in this context is the loss of one or more bases from the DNA so that the 3-base codons get garbled eg. CAT GAC CCT would become ATG ACC CT.......
Slam dunk, right?

A little caution in accepting this story without hard evidence would be wise. In genetics we are taught that frame-shift mutations are extremely disruptive, completely changing the coding sequence and resulting in truncated nonsense. In fact, one term for a frameshift mutation is “nonsense mutation.” A biologist’s basic intuition should be that frameshifts are highly unlikely to produce something useful. The only reasons for the widespread acceptance of Ohno’s hypothesis that I can come up with are the unusual character of the sequence itself, Ohno reputation as a brilliant scientist (which he was), and wish-fulfillment on the part of some evolutionary biologists.

Fortunately, science marches on, and evidence continues to accumulate. The same group of Japanese scientists continued their study of the nylonase genes. nylB appeared to be the result of a gene duplication of nylB′ that occurred some time ago. EII′ (the enzyme encoded by nylB′) had very little nylonase activity, while EII (the enzyme encoded by nylB) was about 1000 fold higher in activity. The two enzymes differed in amino acid sequence at 47 positions out of 392. With some painstaking work, the Japanese determined that just two mutations were sufficient to convert EII′ to the EII level of activity.

It is an interesting story, so read the original link. Note also, that while the story comes from the Discovery Institute, it references the work that resolved this issue.

David
 
Last edited:
It turns out that the Nylonase story is a bit more complicated!


https://www.evolutionnews.org/2017/05/the-nylonase-story-when-imagination-and-facts-collide/




Just to be clear, a frameshift in this context is the loss of one or more bases from the DNA so that the 3-base codons get garbled eg. CAT GAC CCT would become ATG ACC CT.......


It is an interesting story, so read the original link. Note also, that while the story comes from the Discovery Institute, it references the work that resolved this issue.

David
I do not think the discussion about whether or not frame shift is in play negates the value of Nylonase as evidence for recent new function.
This essay:
Is it easy to get a new protein? A reply to Ann Gauger
goes more in to that.

But again you answer a post with several links to good peer reviewed research, with a post on a ID/creationist blog, run by the Discovery Institute.
My main argument in my previous post did not hinge on Nylonase in any way.

If it bothers you, forget about Nylonase, i would rather learn what you think about the rest of my previous post.

Edit: my post missed a later edit of your post.
 
Back
Top