Glancing at Uncommon Descent (I still do as Denyse O’Leary often reports on interesting science articles, as here*, and the odd comment thread can still provide entertainment), I see an OP authored by gpuccio (an Italian medical doctor) entitled The Ubiquitin System: Functional Complexity and Semiosis joined together, telling the story of the ubiquitin protein and its central role in eukaryote biochemistry in some considerable detail. The subtext is that ubiquitin’s role is so widespread and diverse and conserved across all (so far known) eukaryotes, that it defies an evolutionary explanation. This appears to be yet another god-of-the-gaps argument. Who can explain ubiquitin? Take that, evolutionists! I’m not familiar with the ubiquitin system and thank gpuccio for his article (though I did note some similarities to the Wikipedia entry.
In the discussion that follows, gpuccio and others note the lack of response from ID skeptics. Gpuccio remarks:
OK, our interlocutors, as usual, are nowhere to be seen, but at least I have some true friends!
and later:
And contributions from the other side? OK, let’s me count them… Zero?
Well, I can think of a few reasons why the comment thread lacks representatives from “the other side” (presumably those who are in general agreement with mainstream evolutionary biology).
- In a sense, there’s little in gpuccio’s opening post to argue over. It’s a description of a biochemical system first elucidated in the late seventies and into the early eighties. The pioneering work was done by Aaron Ciechanover, Avram Hershko, Irwin Rose (later to win the Nobel prize for chemistry, credited with “the discovery of ubiquitin-mediated protein degradation”, all mainstream scientists.
- Gpuccio hints at the complexity of the system and the “semiotic” aspects. It seems like another god-of-the-gaps argument. Wow, look at the complexity! How could this possibly have evolved! Therefore ID! What might get the attention of science is some theory or hypothesis that could be an alternative, testable explanation for the ubiquitin system. That is not to be found in gpuccio’s OP or subsequent comments.
- Uncommon Descent has an unenviable history on treatment of ID skeptics and their comments. Those who are still able to comment at UD risk the hard work involved in preparing a substantive comment being wasted as comments may never appear or are subsequently deleted and accounts arbitrarily closed.
I’m sure others can add to the list. So I’d like to suggest to gpuccio that he should bring his ideas here if he would like them challenged. If he likes, he can repost his article as an OP here. I guarantee that he (and any other UD regulars who’d like to join in) will be able to participate here without fear of material being deleted or comment privileges being arbitrarily suspended.
Come on, gpuccio. What have you got to lose?
gpuccio@UD
I mean that you are moving goalposts. If you want to argue that the complex function of a single protein cannot be deconstructed into smaller steps, then acquisition of novel protein domains is a valid counterargument. The polite thing to do is not to move on to the discussion of the origin of those domains, until you have acknowledged that your initial claim was defeated.
You don’t mean blind, random, unguided evolution, do you? 🙂
gpuccio@UD
Oh, for Pete’s sake. You are not going to find the handiwork of your Designer with superpowerful bio-informatic tools if you don’t even master statistics 101!
Your data are heteroscedastic because your variable has an upper and lower limit. This happens for all variables with this property, like proportion data: Here is an example I generated from sampling binomial distributions with differing probabilities of success. Within a minute I exactly copied the behaviour of your variable:
colewd,
If you cannot understand every IDist who has attempted to calculate p(T|H) assumes, implicitly, that THIS protein, with THIS degree of constraint, is the ONLY way to achieve [function of interest], then neither I nor my High-School daughter can help you.
Every time an IDist comes along and claims that THIS protein, with THIS degree of constraint, is the ONLY way to achieve [function of interest], subsequent events prove them wrong. OMagain enjoys laughing about “the” bacterial flagellum; John Walker and Praveen Nina laugh about “the” ATPase; Anthony Keefe and Jack Szostak laugh about ATP-binding; now Corneel and I are laughing about ubiquitin ligase: multiple ligases can ubiquinate a given target, therefore the IDist assumption is false. The different ligases that share targets ARE “other peaks”.
This is Texas Sharp Shooter. But now, hilariously, your post-hoc “binding flexibility would be important to this design” is painting a target that resembles Hampton Court Maze.
@ Bill
This is what is wrong with “Islands-of-function” arguments. We don’t know how many needles are in the haystack. G Puccio doesn’t know how many needles are in the haystack. Evolution doesn’t need to search exhaustively, just stumble on a useful needle.
Alan Fox,
This is the very thin thread that evolution is hanging on to prior to falsification.
How can you make this claim after saying you don’t know how many needles are in a haystack that is billions of orders of magnitude larger then the number of elementary particles in the universe?
Unless the needles are almost equal to the haystack the theory is falsified. All the evidence is pointing to needles=haystack being false, especially for nuclear proteins and irreducibly complex structures.
Only if you are sticking your fingers in your ears whenever the evidence is presented to you.
I see DNA_Jock has already picked up on this. All what evidence? Keefe and Szostak did some pioneering work generating random protein samples and testing for just one property, ATP affinity. That didn’t show functionality is rare in sequence space.
What work has been done that demonstrates lack of functionality in proteins as yet unsynthesized and tested for activity? How long would it take to test for all possible activities in all organisms? Lack of functionality is an assumption and work that has been done so far does not support that assumption.
ETA Clarity
Why especially for those things? Do you have something specific you can provide a citation to?
Are you kidding? Then they’d never finish acknowledging that their initial claims have been refuted, and it’s all a landslide from that point on. They’d end up admitting that they have really nothing against evolution but misinformation. Creationism, ahem, I mean, ID wouldn’t survive that kind of politeness. Come on!
These stupid arguments based on kindergarten level premises & math should give pause to anyone with a bunch of functional brain cells. How could the brightest scientists out there have missed such an obvious issue? As a rule of thumb, whenever one finds one of those things convincing, the probability of being right is about 0%. If, on top of that, it’s a widely held belief among creotards, the probability of being wrong is 115%
This is TSZ. We have to dumb it down.
What would?
But first it has to stumble upon a haystack. You make a great argument for design.
Somebody’s got to!
Mung,
If Keefe and Szostak had tried to find binding to a few dozen different ligands and failed every time. THAT would have showed that function is rare in (protein) sequence space.
The reality is that their technique has been used to find peptides that bind to many different ligands. Thus, function is not rare in protein sequence space.
They did fail every time, until they had a success.
ETA: If they had failed every time it would not be rare, it would be non-existent.
Alan Fox,
Just one property, binding ATP, and the haystack is 100 billion times the size of the needles.
It shows it’s incredible rarity. 1 function, 80 AA, and 1 in 100 billion chance of function. How would you think that you could build a flagellum with these odds of getting an 80 AA sequence to perform a single function of binding a small molecule.
What is the haystack in this analogy? Please elaborate.
It generally happens to be the case that in sequence space, new functions are in the vicinity of already existing functions. Most proteins evolved by divergence of duplicate genes rather than de novo.
The individual proteins that make up the flagellum did not need to be “discovered” by sequentially generating billions of random sequences.
Almost all of the proteins of the flagellum are known to have non-flagellum homologous sequences, meaning the flagellum proteins evolved from some ancestral sequence that was closeby in sequence space but performed another function.
This claim has been unfounded every time you’ve made it (and you’ve been making it for years). And every time you’ve been asked to support it, you’ve given articles that failed to support it, in that they didn’t even tangentially touch upon the contention you are being asked to support.
Last time I asked you, you linked some utterly irrelevant article about a tumor suppressor gene. You didn’t even bother with an attempt at an argument for why that tumor suppressor gene constitutes support for your claim.
So I must ask once again, what evidence do you have that nuclear genes are especially rare in sequence space?
Because if all the needles are right next to each other, and an arbitrary one of the needles can be reached from anywhere in the space in less than 20 steps (because the physics of intermolecular bonds), then the total number of needles in the bunch is actually irrelevant.
What matter is the distance from one needle to another, and some times the distance from anywhere to the nearest needle.
Rumraket,
This is your conclusion without evidence to support it other then speculation.
You need to get 30 proteins to bind and to perform a single function. Why would 30 proteins optimized for a different function bind together and perform a new function? Certainly not without modification and Szostacks paper shows 100 billion trials to bind an 80AA protein to a small molecule.
Now you are claiming this happens 30 times in a row with larger molecules. In addition you have to evolve assembly molecules that can build this complex every 20 minutes. So not only do the sequences have to be modified randomly so they can fit and perform a single function they have to be expressed together in the right sequence so they can be built rapidly as a part of cell replication.
Were talking about bacteria at this point 🙂
Is this your standard content-free snark, or are you claiming that they had one or more failures before they tried ATP binding?
No actually, that’s not how math works. If they conducted, say 167 such experiments and failed every time, for a total of 10^15 tries, the most you could say would be that it’s unlikely that function is present at higher than 3 x 10^-15 in protein space. You would not be able to conclude that it’s non-existent, just rare.
I am enjoying colewd being comfortable with his assumption that there’s only one way, when we know that there are multiple ways, and our best estimate, from Keefe and Szostak, is that there’s 10^93 ways.
LOL.
And he thinks that an event that occurs one in 100 billion times is rare.
Hey Bill, in the time it took you to read this post, a 1 in 100 billion event occurred inside YOU.
(100 trillion bacteria, conservative rate of replication every 100 minutes -> one trillion replications per minute…)
Dammit, I need a new meter!
I did not realize that all 30 proteins are necessary to the function of all bacterial flagellae.
Ah the “twas ever thus” fallacy. My favorite. Although at 20 minute doubling times, there’s a one in 100 billion event occurring inside Bill Cole every 1.2 seconds. Yikes!
Rumraket,
The claim is supported by mutational sensitivity to AA change. This is the argument that gpuccio is making by showing evidence of purifying selection. Jock countered with speculation of multiple peaks but without empirical support.
I think you’re misunderstanding a few things.
1. Of course, the mutations that make the “news” are the bad ones. So, we know about the mutations that make a tumor “suppressor” gene fail because, well, they make the gene fail. Obviously, mutations that don’t harm the function never make it to the “news” because they don’t make anybody sick, and thus, we have a lot of “healthy” mutation-space ignored. This is not speculation, this is fact. Furthermore, now that whole genomes are being sequenced, scientists are discovering some pretty spectacular mutations (from point mutations to large pieces of chromosomes missing), that have no effects on the carrier’s health.
2. High purifying selection doesn’t mean no change at all. You should check a few homologous protein families. Quite variable those sequences. All that means is that the tendency is for conservation, but not that no changes occur. Otherwise all proteins in a family would be identical. Guess what? They aren’t. Not by a long shot. What would that variability mean if not that there’s either lots of peaks, or that the peaks are very, but very, wide? Now, I’m inclined to multiple peaks for a reason. If it was a “wide” peak, then there would be less resistance to change. This is not speculation, this is hard data. the variability is evident, the synonymous and non-synonymous substitutions are right there to tell the story of the changes happening, with resistance to change here and there, but change happening regardless. The best interpretation is that the exploration doesn’t always find good variants, but that it finds good variants anyway. Therefore multiple “peaks.”
Make sure you understand it before responding. Please.
I doubt that I’m the first to tell you any of this. It seems like the discussion has been going on and I haven’t had the time to catch up. But these guys are very good and much more patient than me. Maybe you should read their comments a bit more carefully.
colewd:
Rumraket:
Bill,
Instead of wasting your time coming up with idiotic arguments for design, why not read Arrival of the Fittest? It discusses this topic in detail.
You’re unlikely to understand the book on your own, so when you’re feeling confused or baffled, bring your questions here. Someone will help you, especially if you’re making a good faith effort to understand this stuff instead of mindlessly rejecting it.
It’s a much better investment of time than endless cycles of “Bill makes dumb argument; brighter folks refute it”.
Actually, these are two different arguments. The degree of constraint around an optimized peak tells you about the degree of constraint around that optimized peak. This is your “mutational sensitivity”. And the constraint is a lot less tight than you are implying. And it tells you virtually nothing about the degree of constraint in less optimized sequences. Read Hayashi. Then there’s a separate argument about the existence of “other peaks”. For which there is abundant empirical support. We’ve given you examples for ATP binding, ATP synthesis, ubiquitin ligase, and bacterial flagellae.
keiths,
I have read the book twice the last time a couple of years ago. If you think it provides a strong counter argument to gpuccio start an op.
DNA_Jock,
Great. Create an op and show support for this claim. Please define what a peak is in your op and how mutation will safely move you from one peak to another. The problem I see is that all we see now is the DNA of extant organisms. If I am right the peak discussion is speculation by definition and based on circular reasoning.
colewd:
As I said:
It’s clear that you didn’t understand the book. If you had, you wouldn’t be making dumb statements like this:
What was it that confused you?
As I have already stated (twice), we’ve given you examples for ATP binding (10^93 peaks…), ATP synthesis, ubiquitin ligase, and bacterial flagellae. It’s a pretty uncontroversial claim, Bill, and one to which you haven’t offered any counter-argument, beyond personal incredulity. Not really OP-worthy.
It won’t, if the landscape is invariant. What a silly thing to write. Do you even understand what a peak is? Oh! I’m sorry, here goes: a peak is a contiguous region with fitness above a given threshold.
Based solely on the consequent there, I think it is safe to say that you are wrong.
Dear admins, it seems that my comments have become invisible to the good people at UD. Could this technical inconvenience be solved please?
That’s idiotic. Those proteins with different functions exist, and they are homologous to the flagellum proteins. It’s not speculation, it’s a fact.
They already do.
There are also “versions” of the flagellum known where key structures are missing, yet the whole structure retains another function than motility.
To give an example, in the symbiotic bacteria Buchnera, the bacterial flagellum is missing all proteins associated with filament assembly and the filament itself. It is simply gone, not there. The filament as you will recall is the structure that is analogous to a propeller, which in the flagellum is the long tail that spins in the extracellular environment and generates movement. The entire rest of the flagellum (with the exception of the stator) is still present, but it still functions as a membrane protein translocase.
In the views of some ID proponents, a favorite protein of in the complete flagellum structure is the hook, which in the complete flagellum transfers torque from the axle to the filament. They argue this protein has no other function than the transfer of torque and so would be useless in a reduced flagellum.
But the hook exists in the Buchnera flagellum and is part of the membrane protein translocase. So it’s function is to assist in transporting proteins across the membrane.
The same function that scientists initially speculated (for example in Matzke’s 2003 article) was the ancestral function of whatever pre-flagellum structure from which the flagellum evolved.
See this is your mistake, you think the 30 proteins have to arise de novo and come together in some giant fluke, rather than just slowly changing one or a few at a time from roles in similar structures with closely related functions.
The functions of the other proteins are all very closely related to what they are doing in the flagellum. Structural support, transporting ions or other proteins across membranes, etc.
The filament, for example, is homologous to extracellular adhesion proteins.
Again, the flagellum proteins would not have to evolve de novo, as almost all of them are known to be homologous to other proteins, so would not have to be generated completely from scratch.
No, I’m not suggesting they evolved de novo 30 times in parallel. Actually I’m suggesting they adapted mostly in parallel from their roles in a closely related structure.
Which would of course all already exist otherwise the structures from which the flagellum evolved could not themselves exist.
Which they would already do.
And what’s is the problem again?
How does that support the claim that nuclear proteins are especially rare in protein sequence space? Please spell that one out.
It doesn’t make it any less nonsensical that gpuccio is making it too.
There are highly conserved proteins outside the nucleus, and highly conserved proteins inside the nucleus, just as there are much less conserved proteins outside the nucleus and much less conserved proteins inside the nucleus.
But even if you could show that nuclear proteins were generally more conserved than cytoplasmic or extracellular proteins, that STILL wouldn’t mean that functional nuclear proteins are more rare in sequence space.
Bill, in most cases where scientists evolve de novo a protein with a known biological function (such as an enzyme with esterase activity, or binding to another protein or small molecules), the proteins they evolve have no homology to the proteins that exist in living organisms. That means an entirely new fold that can do the same function as the enzymes and other proteins that exist in life as we know it, has been found. Those are different peaks in the fitness landscape of those protein functions.
For example the ATP binding protein evolved by Keefe & Szostak has no homology to ATP binding proteins from life, yet it works. The minor coat protein from Hayashi et al has no homology to the D2 domain from the bacteriophage, yet can do the same function as the D2 domain. There are many similar experiments.
I decided to take a look at the gpuccio’s responses to the Szostak paper he points to here https://uncommondescent.com/intelligent-design/the-ubiquitin-system-functional-complexity-and-semiosis-joined-together/#comment-656117
Here’s the first one:
Holy fucking shit, I’m speechless. How come no evolutionist thought about declaring every possible sequence as functional based on some arbitrary definition? We could have used that powerful tactic to fool people into buying the darwinian hoax by the millions.
Wait, did I say that out loud? Damn it
gpuccio@UD
gpuccio, all that you wrote withstanding, you are committing the lottery fallacy.
Although he didn’t realise it, it was Origenes that correctly identified what was wrong with your argument:
Indeed, and your specification is simply not independent, is it? It is, as pointed out by DNA_Jock, a fine example of Texas Sharpshooting. This leads you to specify an unreasonably small target space.
Think about it: it was you who suggested that the ubiquitin system has semiotic structure. In how many ways can a biological system tag a protein for degradation by covalently binding a small signalling molecule? Is there one way? Off the top of my head, I can think of several systems of posttranslational modification in extant systems alone: methylation, acetylation, phosphorylation, glycation and ubiquitination. There must be dozens more, and there must be literally BILLIONS of possible ways of accomplishing the very same thing.
You really are the guys that get dealt two pairs in a regular game of Poker and then marvel at the incredibly slim chance of receiving two Jacks and two fives.
gpuccio reminds me a lot of fifth, in that he just can’t shake his assumptions and more importantly, he can’t keep himself from imposing them onto others’ views: it’s all teleological, it just must be, whatever we see it must have been prespecified, so if any mechanism must work, it must produce those “prespecified” targets. Oh, evolution/darwinism has no targets? Cool! then it doesn’t work… by definition!
Dammit, Corneel, I was saving that Origenes quote for later.
😀
In an ironic way there is some truth to Gpuccio’s comment regarding the definition of functoin, in that function really is context-specific.
If you were to somehow replace my lungs with gills, I’d asphyxiate and die. Does that mean gills are nonfunctoinal? No, it just means that what functions under some conditions will not function under other conditions. There is probably no such thing as an unconditional function, it will always be specific to a particular circumstance and environment.
All the bacteria that live in my gut would die in the environments where hyperthermophiles that live inside hydrothermal vents live. All their proteins would unfold, denature and quickly hydrolyze. Does that then mean that their proteins are nonfunctional? No, it just means they only perform the functions they do under certain conditions. We technically can’t claim that they are actually nonfunctional under all conditions.
The reverse might also be true, in that some of these hyperthermophiles go dormant under the temperatures where for example E coli thrives. The activities of enzymes might fall below detectable levels, or stop working entirely if it gets too cold. But could those enzymes be converted through mutation and selection to work at lower temperatures, or higher temperatures? Probably.
I’ve read that ribosomes encoded by the mitochondrial genome are significantly smaller than nuclear encoded ribosomes, and that a much larger fraction of their mass is constituted of protein, rather than constituted of ribosomal RNA as is generally the case for bacterial and eukaryotic nuclear encoded ribosomes. That seems to imply that a large portion of ribosomal RNA actually isn’t critical to the function of the ribosome, can be dispensed with, and a portion of the remainder can can be replaced by protein (and the reverse would be true of course), but that such an exchance is only selectively beneficial depending on the environmental circumstances.
They should get their DNA tested at UD. I have good reason to believe they are self-awareness-NULL mutants, 😀
You should read the papers by the Thornton lab where they use ancestral sequence reconstruction to determine how different proteins have changed functions and specificities over geological time. And how in fact they could have evolved by alternative routes in addition to the ones that happened to occur.
See for example:
Starr TN, Picton LK, Thornton JW.(2017) Alternative evolutionary histories in the sequence space of an ancient protein. Nature. 2017 Sep 21;549(7672):409-413. [doi: 10.1038/nature23902]
I’ve been busy. But thanks for the reminder.
gpuccio, you gave your 500-bit threshold as a figure which is “the foundation itself of ID”. Then you said that it had to arise at once in one mutation.
Now it is extremely clear in William Dembski’s works that the target region (whose probability in the initial monkeys-with-typewriters distribution is less than $2^{-500}$) is supposed to have been shown by him to be unreachable by natural selection. That is clear from the way he tried to put forward a Law of Conservation of Complex Specified Information. Most of chapter 4 of his 2002 book, reprinted in 2007, was devoted to that argument.
To be reachable by natural selection, one does not just ask if one can get all the way in one new mutation. One also asks about cumulative changes of smaller size.
You may say that you aren’t talking about Dembski’s argument, but I remind you that you said that this 500-bit figure is “the foundation itself of ID”. Did you mean to exclude Dembski’s 500-bit figure from that? After all, his work is where it comes from.
My views aren’t “strange”. I have closely read Dembski and tried to figure out his logic, using the most generous interpretation I can make. When I have mistaken his views, I have admitted that and corrected my argument.
I am not seeing a problem with our software. Maybe it’s a problem with their wetware.
Sorry Neil. I already came to the same diagnosis and my request was meant to be read tongue-in-cheek. Thanks anyway.
I guessed that, but decided to reply anyway (as a formality).
Rumraket,
Evolution can go in any direction except once it starts down a path it is committed to that path as living organisms are made of interdependent parts. The mismatch you are describing sucks 🙂
BTW: Thanks for the Thornton abstract. I have read some of his previous papers.
Do you define a “peak” as a new function?
That just supports my argument. Functions are defined and dependent on a context. There is no such thing as an unconditional function.
Any function can have one or more peaks, or hills if you will. A peak is just an area in the sequence space where a collection of similar sequences that perform the function, exists.
Some hills are so broad that the sequences that make up that hill can be completely dissimilar to each other. Others are more like tiny spikes. If moving from one functional sequence to another performing the same function cannot avoid going through a valley of nonfunctionality, then at least two hills must exist with that function.
But what you can see, if you understand that the existence of a hill depends on a particular context, is that the fitness landscape of protein functions in sequence space, is dynamic, not static.
In a dynamic fitness landscape the topology of the fitness landscape can change as the environment changes.
For example, if you gradually start increasing the temperature, hills will start to disappear as the proteins that make up those hills would denature and no longer be able to perform the function.
In this view of the fitness landscape of protein functions in sequence space, it is like the surface of a billowing sea. Peaks and valleys come and go as waves roll across the surface.
The environment (which here can include anything from the presence of other genes, to the physics and chemistry of the intra and extracellular environment) changes over deep time. Mutations is the process that moves protein sequences around on the surface of this sea, and selection is the tendency for that movement to be towards the tops of the waves.