It has been widely advertised that nylon eating genes evolved after 1940. I have no problem with that claim in principle since new antibiotic and malaria resistances have evolved since 1940. Even though I can easily accept the possibility of post-1940 nylon-eating evolution in principle, where is the slam dunk evidence that this is actually the case? Did a significant portion of the ability for bacteria to digest nylon take place after 1940 (or 1935 when nylon was first created)?
In an NCSE article New Proteins Without God’s Help we read:
Since nylon first came into commercial production in 1940, we know that the new enzymes have formed since that time.
A similar line of thinking was argued at BioLogos in Biological Information and Intelligent Design, De Novo or Ex Nihilo.
Three nylon eating genes NylA, NylB, NylC were discovered on the Flavobacteria plasmid pOAD2 from 1977-1992, but the researchers concede none of the three have significant sequence homology. Worse, in papers published in 2007, they reported other bacteria contain those same genes in their chromosomes. Unless the researchers have access to pre-1935 bacteria sitting in lab refrigerators, the claim that the genes actually evolved new proteins via mutation is dubious since we have no pre-1935 bacterial samples to actually do a comparison with, especially in the case of NylC. The claim that NylB’s nylon eating ability evolved by gene duplication from a non-functional NylB-prime gene could just as well be interpreted that a functionless NylB-prime gene is a defective copy of a functioning NylB gene!
What’s the proof new nylon eating genes actually evolved after 1940, or is it just speculation? Slam dunk proof would entail having strains of pre-1935 bacteria and then comparing it with the strains after 1935 that supposedly evolved new genes. Is that the case? No. Just speculation which began in 1977 but got less defensible over the next 40 years as more bacteria and non-sequence-homologous genes were discovered to have nylon eating capability.
At least we can credit Richard Lenski who can back up his claims of evolution because he has samples of bacteria in his lab before and after his creatures evolved, whereas the guys promoting the claim the nylon-eaters are new don’t have pre-1935 physical samples of the bacteria. The only pre-1935 samples of the bacteria they have are samples from their imagination.
Incidentally, it is worth pointing out the bacteria that is the centerpiece of the controversy has been mis-classified or misidentified, and thus the name of the bacteria has itself gone through some sort of macro evolution. It was first called Acromobacter gattatus, then renamed as Flavobacterium sp., but then renamed Arthrobacter sp. after it was realized it had been put in the wrong taxonomic category all along.
But anyway, here is the 1977 speculation that became uncritically accepted and exaggerated as fact for the last 40 years:
There are two possible mechanisms for an enzyme to be active towards unnatural synthetic substance such as 6-aminohexanoic acid cyclic dimer; one is that an unnatural compound is hydrolyzed as an analogue of the physiological substrate, and the other is that the compound is hydrolyzed by an evolved enzyme which originally had an activity on a physiological substrate but lost it by the evolutionary mutation. The data obtained in this experiment indicate that this enzyme did not hydrolyze any physiological substrates tested including peptides, cyclic amides, and amides. In addition to the lack of activity on natural compounds, its low turnover number (8 SKcI) compared to other cyclic amide hydrolases (35 – 260 s ~ for penicillinase [12]) supports the possibility that the enzyme has evolved by adaptation to a new synthetic substance which is a waste product of nylon-6 production.
I have no problem if they say “the enzyme has evolved by adaptation to a new synthetic substance which is a waste product of nylon-6 production” when they mean an existing enzyme has changed its pre-existing function to a different one. That is to say, it acquires a new function while also losing its previous one. However, I would have serious issue with the NCSE’s insinuation that this was such a spectacular change that it illustrates how to solve the origin things on the order of the spliceosome or ribosome.
They don’t have strains sitting in a refrigerator from which to compare the “newly evolved” genes with do they? They just guessed, and then 30 years later one of the same team (Negoro) found nylon eating genes on other bacteria, plus reasonable nylon-eating homologues on other bacteria to boot.
Worse, the claim that these are post-1935 genes is weakened by some facts that require a little elaboration. First there is the basic nylon monomer that forms the components of nylon oligomers. NylA, NylB, NylC have different specialties of which nylon oligomer they can degrade. These three enzymes provide what looks like a digestive cascade where N-oligomer substrates are degraded by NylC to dimers which can be further degraded by NylB (and even NylC but to lesser extent). The cascade is described in: Biodegradation of nylon oligomers.
But important to note: NylA, NylB, NylC are not sequence homologous. In A New Nylon Oligomer Degradation Gene (nylC) on Plasmid pOAD2 from a Flavobactenium sp we read:
Sequence alignment by the method of Wirbur and Lipman (21) showed no significant homology among the
nylA, nylB, and nylC genes. These results suggest that the three nylon oligomer-degradative enzymes evolved independently.
There is no pre-1935 bacteria from which to compare with to establish NylC’s true history, much less anything beyond speculation about the nature of NylC’s ancestor. So how can anyone claim it is new (as in post-1935-new)? Just guesses. They’d have to argue these the genes evolved independently since 1935, and in the case of NylC, the ancestor is almost totally without detailed description.
They however argue, a functional NylB arose from a non-functioning NylB-prime gene via gene duplication in a 1991 paper entitled: Evolutionary adaptation of plasmid-encoded enzymes for degrading nylon oligomers.
But for all we know NylB’ is a defective copy of a functioning NylB! Where is a Ken-Miller-pseudo-gene-like argument when you need it?
In fairness, another bacteria called Pseudomonas was shown in lab conditions to acquire enzymatic nylon-eating function after a few months. The gene however in Pseudomonas, according to BLASTN, has no sequence homology with Flavobacteria. However, the protein according to BLASTP can align 91% at 37% homology. But even the researchers were astonishingly vague about what actually changed in the bacteria that enabled nylon degredation. The two relevant papers are: Characterization of the 6-aminohexanoate-dimer hydrolase from Pseudomonas sp. NK87 and Emergence of Nylon Oligomer Degradation Enzymes in Pseudomonas aeruginosa PAO through Experimental Evolution.
Unless we have strains sitting in a refrigerator of pre-1935 bacteria, how do we know that the genes actually evolved to digest bacteria? For all we know, it had that capability already or very close to it. And unless we have genes from all existing bacteria in 1935 and can prove all the bacteria on the globe do not have horizontally-transferrable nylon eating genes which could end up on the plasmids of Flavobacteria, we can’t unequivocally argue the three nylon eating genes (NylA, NylB, NylC) in Flavobacteria didn’t already exist pre-1935.
Even if the capability of nylon eating evolved, it is also not necessarily because of a new gene, but new regulation. A mini-example of this possibility had to be raised because even in their lab, they had bacteria with identical nylon eating genes but which could not digest nylon.
it is conceivable that expression of the nylC gene is enhanced in these strains and the elevated enzyme activities made the cells Nom+. However, the following possibilities could not be ignored: (i) KI725R strains may possess an additional nylon oligomer-degrading enzyme which is active
toward a substrate included in Noml, but K1725 has no degradative ability toward the substrate; (ii) nylon oligomer transport proteins are activated in K1725R strains; or (iii) the EIII proteins were altered by mutations in the coding region of the nylC gene by which the specific activities and/or
substrate specificity of the enzyme were changed. The last possibility should be negligible,
So my question is, if we ignore the NCSE hype, does the scientific literature really even prove unequivocally a significant part of nylon eating genes evolved after 1935 and absolutely rule out these genes pre-existed? Maybe yes, maybe no, but the NCSE can’t now brag that any significant evolution (like say compared to evolution of antibiotic resistance) bacteria definitely evolved after 1940 without God’s help. We simply don’t know.
NOTES:
Here are some accession numbers:
NylB (protein) in Flavobacteria: WP_012476894.1
which generates a BLASTP hit on Agromyces accession number: BAE97621.1
NylC (protein) in Flavobacteria: BAA01528
which generates a BLASTP hit also on Agromyces accession number: BAE97629.1
in Pseudomonas, the nylon eating gene accession numbers are : BAA01524 (protein), D10678.1 (DNA)
which generates BLAST hits on a variety of organisms as well
You mean the part that doesn’t even need Ohno’s frame shift? like:
🙂
Of course, that’s the point, the ancestral protein extends beyond the one created by the frameshift mutation, in both directions.
Frameshift mutations can happen in the beginning of the previous reading frame, but there’s be no reason to expect that over any other particular location.
So in so far as frameshift mutations take place, there’s often going to be some remnant of the previous coding region upstream (and/or downstream) from the reading frame created by the frameshift mutation.
That’s exactly what we find. There’s a coding region there in the DNA, and it starts properly with a start-codon and a Shine-Dalgarno sequence. And predicts a stretch of amino acids known to exist in other proteins.
First of all, I’m not claiming that is what happened. You seem to have this habit of thinking that homologous sequences from other species must have come from the pOAD2 plasmid, and it must have happened somewhere in the 20th century. At this stage I’m not convinced anything like this must have happened to explain what we see.
Second, the sequence is at the c-terminal (the end), not in the middle of the Nitrate Reductase. So if it got horizontally transferred we might be looking at a gene-fusion of some sort.
The DNA sequence for the pOAD2 plasmid from Agromyces, is highly similar to the same DNA sequence from pseudomonas (as you can see in Okata H et al 1983). This implies this stretch has been under purifying selection in both species, ever since the time when P-NylB and F-NylB diverged. Probably because it has been serving as binding spots for transcription and translation factors that initiate transcription and translation of the NylB, or other downstream genes on the plasmid.
So technically, the putative ancestral 427 amino acid gene could have been transferred to Leucobacter all the way back from before P-NylB and F-NylB diverged, and a gene-fusion happened at some point along the way.
We’ve already been over this. Ohno is surely mistaken about when the frameshift mutation must have happened. As John Harshman explained on the beginning page of this thread.
Thank you for re-stating all the same things over again that we all previously agreed upon. His inferred timing of the frameshift mutation is wrong. He’s not wrong about there being actual evidence that a frameshift mutation happened.
What a truly unsettling revelation for me who is already convinced by the available evidence that life has existed and diverged for over 3.8 billion years.
We settled this crap months ago. Now you’re back as if you’ve made the discovery of the century. LOL
The million year implied divergence of P-NylB and F-NylB is a problem for you and your young-Earth-creationism, not for me.
The bolded part you got right. 🙂
Yeah, why does that sequence even code for protein and why is it matching 97% a fragment of equal size the c-terminal from nitrate reductase in Leubacter?
Why do we find homologous sequences for Ohno’s inferred alternative reading frame that overlaps NylB, if there was never a real protein with that inferred amino acid sequence?
All of these things must be colossal mysteries and coincidences your “creator” put there to confuse us. Or what? I’m sure you can come up with some sort of ridiculous ad-hoc rationalization for all of it. God is an artist maybe. Or maybe God is “helping us understand ourselves by making other organisms be similar in completely nonsensical ways” or whatever similar crap it was you made up last time. LOL
Yes yes, okay Sal. Halleluja!
Look, in all seriousness, you do have some case to make for a failure of scholarship regarding Ohno’s inferred timing of the events, and the subsequent reporting of the events. All of that is perfectly legitimate. What you seem to have a hard time accepting is that there is still evidence that a frameshift mutation happened, and that in this sense Ohno wasn’t wrong.
Play the scholarship angle, forget trying to prove your young Earth creationism, because the evidence isn’t consistent with it regardless of the mistaken inferences by Ohno and the associated problems with peer review.
Rumraket advises Sal:
You’re going to make the baby — and the baby Jesus — cry, Rumraket.
Rumraket,
Thanks for engaging in the debate in ernest with Sal. What do you think is the strongest evidence, in this case, that an ancestral frameshift mutation successfully resulted in a de novo protein?
The fact that upstream of the NylB reading frame, but shifted by a single nucleotide, there sits another reading frame coding for 33 amino acids (found in another functional protein), with the last amino acid overlapping the start-codon for NylB, combined with the alternative reading frame which overlaps NylB, predicts a protein sequence with a significant similarity to lots of other proteins reported in uniprot. While Ohno himself didn’t have this evidence at the time he made the inference, we now do.
Back on the previous page, Sal correctly reasoned that if “relics” of this purported ancestral sequence could be found, this would have substantiated Ohno’s original inference. In effect, Sal tried to use this as a criterion for falsification of Ohno when he failed to find these “relics”. So if not finding them is evidence against, then finding them is evidence for.
How dare you? I’ve been parking bikes in parking racks for millions of years
Let’s highlight that part…
Not finding something is not evidence for its nonexistence.
It depends on the situation, it certainly can be. Not finding an elephant in my apartment really is evidence there isn’t one there.
If it’s expected to be found when searched for, then not finding it strongly implies it isn’t there. Given how big an elephant is, and how small my apartment is, we would strongly expect to find an elephant if there was one. Of course things are rarely that simple and in reality there’s almost always going to be some confounding factors that can explain why we don’t find what we expect to find.
Some times, absense of evidence actually really is evidence of absense. In the particular case with the frameshift mutation, not finding any relics really would be a bit of evidence against a frameshift mutation, it’s just that it isn’t the only kind of evidence we would expect, so it would not allow us to totally rule out the inference of a frameshift mutation on that alone.
Some times you can be in a position where you have evidence both for and against some hypothesis, and in that case it might still be the case that the evidence on one side heavily outweighs the evidence against it. All of this is textbook Bayesian reasoning afaik.
Rumraket,
How close to exhaustive is the current sequence database please?
Relics of the frame-shift means one should disregard portions that are not shifted like:
MGYIDLSAPVAMIVSGGLYYLFTRRGYTFGDTR
which is problematic for this boast:
Yes, for the part that isn’t subject to the frame-shift! You’ll get that even without the frame shift as I demonstrated.
Furthermore, the other search done such as:
Residues are not DNA bases and therefore are not an accurate test of the frame shift, not to mention you didn’t state how much coverage was involved of the 394 residues. If it’s a fragment and not all of the 394 residues, it’s not that great, not to mention, do you want to stake your claim on an E value of e^-22?
You all seem to complain that Dembski’s 10^150 is not strong enough, but you’ll slobber over e^22 which is about 3.58×10^9.
I see where you are going with this and that’s a good point of course. A few years ago I read an estimate that so far, we have sequences from less then 0.3% of the diversity of life.
This problem is exacerbated further by the fact that when we (for example) search the uniprot database which afaik at the present time contains over 12 million protein sequences with a blast-search for similar sequences, it doesn’t actually go through all those 12 million proteins and try to align the sequence you search with, to every one. That is simply not computationally sensible at the present time given how many people could be searching simultaneously at any given moment.
So necessarily the algorithm must be taking some shortcuts, which in turn means there is some chance it’s going to miss something that could be significant. And then there’s the fact that with the default search settings, it might even find alignments, but discard them because they’re judged not be sufficiently significant.
I can’t tell you in quantitative terms how thorough a default search is, you’d have to talk to someone who knows more about that stuff.
Rumraket,
Thanks Rum.
I take it that 0.3% means 0.3% of the estimated total number of species have been sequenced, right?
I’m guessing it’s impossible to know how exhaustive the database is with regards to genes though?.
They are shifted. The codon for R in that sequence overlaps the start-codon for NylB by one nucleotide. So it’ is out of frame of NylB.
No Sal, wrong. I used the amino acid sequence that you yourself provided here, and then I deliberately excluded the upstream-from-NylB region (in order to exclude hits for that part and focus entirely on the reading frame that overlaps but is shifted from the NylB coding region).
So what remains is the part that overlaps the NylB coding region, but at a reading frame that is shifted.
Look, you are the one that bothered to compile the amino acid sequence that Ohno inferred from the alternative reading frame overlapping NylB, and you looked for “relics” of it.
I argued why failing to find such relics would not be sufficient to claim that there was no evidence that a frameshift mutation happened, by simply using the same inferences Ohno made (which were based on just looking at the DNA of the NylB gene and a smaller portion of the up and dowstream regions).
But it really would serve to further substantiate Ohno’s case if those relics of the actual amino acid sequence could be found.
The total length of the queried section is 373 amino acids. With some gaps, this is what aligns to a 394 length piece of a larger protein with a 36.3% identity score and an E-value of 1.1×10^-22. I provided links for these alignments so you can check them yourself.
Apparently this 390 ish segment is part of a much larger 4440 amino acid protein. But Sal, look at the link. Notice how this same segment aligns one after another over ten times in a row. There’s a nice list of alignments there you can scroll through, and it even shows with a nice graphic where the alignment sits in the bigger 4440 amino acid protein.
That means this huge 4440 amino acid protein is apparently a big repeat of the same structure. Probably created by multiple gene-duplications. Apparently it is a multienzyme complex that synthesizes a range of related antibiotics. And this multienzyme complex has evolved by duplication and subfunctionalization.
Uhh no, 1.1e-22 means 1.1×10^-22.
Regardless, this is just a flailing nonsensical distraction from you. The problem isn’t that Demski’s number isn’t “strong enough”, it’s that it’s a nonsensical calculation.
That is how I understood it, yes.
Well, looking at the rate of discovery of truly novel protein folds, this seems to have dropped off considerably in recent years, implying that despite the low overall diversity sampling, most of life on Earth appears to be mostly re-using the same narrower set of protein modules. Which is explained by common descent of course.
Thanks again, I asked because I remember reading there are just over a thousand distinct protein folds (something like 1300-1400 IIRC?) in all life diversity. Of course not all genes code for proteins, but since the topic at hand does involve a protein sequence, it seems to me it’s safe to say that we have a rather exhaustive library of proteins, and therefore, failure to find homology in a complete database search would have pretty much proved Sal was right. Even a non exhaustive search (given the limitations you mentioned above) but with enough coverage would have been strong evidence that Sal was right.
Clueless as I am about even the basics of molecular biology, it seems to me that the fact that those relic homologs do actually exist, it follows that you do in fact own a brand new bike parking rack, and Sal’s failure to find those homologs is only a testament of his own failure
OK how about this, I deleted the “R” at the end.
MGYIDLSAPVAMIVSGGLYYLFTRRGYTFGDT
Which looks pretty much the same. 🙂
It should still give hundreds of hits. This is the unshifted portion. Since it is the unshifted portion, how does this lend credence to Ohno’s shift hypothesis?
Nothing requires that short sequence be functional to end up somewhere according to evolutionary theory. That’s consistent with the whole shpeel of taking random junk and making something of it.
So far as I can tell, there are no relevant unequivocal relics. For now Ohno’s shift hypothesis looks like a pile of shift.
I’m going to give you a response to this, but before I do, I want you to go buy stocks in asphalt, because there’s going to be space for a megamall parking lot between your asscheeks.
I will begin by giving you this small hint:
I took the full amino acid sequence of Ohno’s inferred ancestral protein that you yourself supplied (including both the upstream region that does not overlap the NylB gene, and the part that does), and sent it to the I-TASSER server so they could try to model the structure of this protein.
I’ve been waiting for the results these last few days, and when I logged onto my email this morning they arrived.
I-TASSER comes with this handy feature that when you send an amino acid sequence for structural modeling, once a model has been generated, the server searches the databases for similar structures from known functional proteins.
I think you can see where this is going. I will post the results later. Right now, I just want to take pleasure in them myself. And I recommend you invest in asphalt. There will be more excess space for parking than at the creationist museum.
HAHAHA.
There will be nice pictures that show the structural match. So when Sal uses the words “relevant unequivocal relics”, it will be hilarious to see him try to spin results we can see with our own eyes.
Here’s what I wrote wednesday:
While wondering about that, I decided to do that work myself and submitted the sequence Sal was kind enough to compile for us, to I-TASSER.
This is too funny.
So how does this prove the specific DNA sequence?
Structurally similar proteins are often only 12% sequence similar.
All you may end up proving is convergence on the same folds, something you advocated a lot in this forum. Yeah, wasn’t Allen Miller who said there or only a limited number of folds.
So how then does this prove this Ohno’s DNA sequence actually existed. It doesn’t. Just a rabbit trail for someone grabbing at straws.
And I must point out, you haven’t made the retraction regarding this amino acid sequence that comes from an unshifted part of RS-IIA:
MGYIDLSAPVAMIVSGGLYYLFTRRGYTFGDT
which you used to argue for Ohno’s frame shift. At least I’d think you’d resort trying to find similarities to the shifted portion rather than the unshifted portion if you’re going to try to establish the frame shift happened.
How about you pump some random DNA sequences the create open reading frames of sufficient length and see if the predicted proteins converges on existing folds. You and Allen seem to think it will. I actually think that it might.
But convergence on a fold predicts that a DNA sequence actually had an insertion event how?
Rumraket,
Btw, be sure to correct for Ohno’s typo so you get an even longer reading frame and more honest results. I mean, what if that protein should actually be a thousand residues long and you also find it structurally similar to something. That would be a bit embarrassing to your thesis.
Sal, keep posting. 🙂 Haha oh man.
I will take this as a total concession from you. The entire creationist case against the evolution of new functional protein folds is now in wrecks. You’ve conceded the entire argument.
I didn’t even have to post my results, and here you are having utterly capitulated in effect.
I accept your capitulation.
Personally, I’m not so sure folds are that unlikely. I’m undecided.
But, let me point out your position isn’t as strong as you think.
If you think it’s not that hard to have random sequence arrive at a fold, then it’s not that hard that Ohno’s PR.C arrives at a fold, therefore it doesn’t make Ohno’s sequence especially likely to be better than random, which reduces the likelihood he was right because you need Ohno’s frame shifted sequence to be special. Therefore in attempting to get folds to be easily achieved, you disprove the specialness of Ohno’s sequence, which means then there isn’t much need to invoke the frame shift.
Btw, you still have acknowledged your gaff here where you cited the unshifted portion of Ohno’s sequence as evidence of a shift. Do you really not understand the problem with your line of reasoning:
And which results did you examine, the one with Ohno’s typo or the one without? You don’t want an inadvertent deletion due to Ohno’s typo as that will cause another frame shift, and you of all people should really care about that since you swear by shifts. You swear by shifts so much, you’re just like Ohno, full of shift.
Stick to the facts, Sal, please. 🙂
Alrightly.
To Rumraket, there are two sequences to examine, one with Ohno’s typo, the one I posted, and then the one without.
I don’t think it would help your position to be saying the sequences with and without Ohno’s typo both arrive at a some homology. Ohno didn’t justify why he made that typo. If a random typo can create a fold with I-TASSER then something is problematic.
But I already stated my position, I’m undecided about fold improbability, at least what our computers are able to deduce since the protein folding problem is formally unsolved.
I alluded to the issue: many structurally similar proteins have only 12% (sometimes less) sequence identity.
I’m not like other creationists in that regard that I think folds are necessarily improbable. I’ll cite Axe’s work as to put the issue on the table, but with some reluctance.
Anyhow, I got what I really needed from this discussion. I bold the parts you me and John Harshman agree on.
and
So if compose and essay about why nylonase ability didn’t evolve via frameshift post-1935, I shouldn’t get any disagreements. We’ll disagree about other things, but I think I’ve defended that part of the issues sufficiently well at this point to not get any material disagreement if I said,
Ohno’s shift hypothesis entails changing in the range of 392-427 residues instantly. I’ve gone on record and said, citing Kato 1991, as little as a 2-residue change could do the trick, maybe not even that because we recently found bacteria from the Indian Ocean, which is not exactly a nylon rich environment, which could eat nylon. We have predicted nylonase genes in artic bacteria. Heck, we humans have trypsin, and that is a nylonase!
Note, having a nylonase enzyme doesn’t mean the creature can digest nylon. In fact, unless one can digest the breakdown products of nylon, namely, 6-aminohexanoic acid which is aminocaproic acid, which by the way is a pharmaceutical known as AMICAR, the result could be toxic.
https://www.drugs.com/pro/amicar.html
Some of my creationists colleagues might be alarmed that I’m not decided over whether folds are that improbable.
But protein function isn’t all about folds!
Here is an example of unfolded tails of histones. I don’t think anyone believes we can willy nilly monkey with the sequences of these unfolded tails as the tails implement the histone code in chromatin:
There are certain caveats. I’d have to check the original paper to see if Ohno actually claimed a post-1935 frame shift or that the frame shift resulted directly in nylonase. Otherwise, the claim you make above is much weaker than is warranted by the data. It isn’t just unproven or implausible. It’s an impossible conclusion from the data, or at least as impossible as anything is in science. The mere existence of nylB’ is enough to tell us that.
This seems utterly wrong to me
I believe “not that hard” is taken to mean “there are plenty sequences in the vicinity of a folding protein to enable evolution”, not that most random sequences will fold just fine
AM I missing something?
I don’t follow your argument here.
This makes no sense whatsoever. I don’t care about “specialness”, or how “much need” there is to invoke a frameshift mutation.
What I care about at this stage is if Ohno was actually right, that there really was another, longer reading frame, on the pOAD2 plasmid coding for a functional protein, and did the F-NylB and P-NylB common ancestor emerge by frameshift mutation from this putative ancestral gene.
Whether frameshift mutations resulting in new functional protein folds emerging by chance, is a rare and “special” event or not is not really interesting to me. It isn’t necessary for any other beliefs I hold that Ohno was really right. We started arguing about this because you insinuated that Ohno had nothing to go on, and I disagreed and I think his original reasoning was actually pretty good regarding there having been a frameshift mutation.
This thing with the alignments I found is just a big bonus. If Ohno was actually wrong and there never was a frameshift mutation, then okay, Ohno made an inference that didn’t stand the test of time.
But it turns out that I’ve found some results that, as I understand them, imply that Ohno was right in there being an original longer reading frame, and that the NylB ancestor emerged inside this longer reading frame by a frameshift mutation.
That’s not a gaff at all. I’ve explained this in detail already. I will try again.
Ohno infers his frameshift mutation happened inside another reading frame.
Like this:
[OLD PR[NEW PROTEIN]IN]
Pretty crappy analogy I admit, but the main point is this. There should be evidence of a piece of protein coding sequence upstream to the [NEW PROTEIN] reading frame. There is, that is what is the part called “[OLD PR” in my analogy here.
This “[OLD PR” part is the sequence you are now telling me is somehow a “gaff” and doesn’t constitute evidence for Ohno’s frameshift mutation.
Of course it does, how can it not?
Not only do we have this “[OLD PR” part for Ohno’s inferred ancestor reading frame, it is also out of frame by a single nucleotide, to the NylB coding region which he argues is what was created by a T-insertion.
So. The codon coding for the last amino acid in that sequence overlaps the NylB reading frame in the start-codon for NylB, but only by a single amino acid. As such, it is first of all not in frame with the NylB frame, so this is exactly what you would expect there to be if a smaller reading frame emerged by a frameshift mutation in a larger reading frame. You would, if that happened, expect there to be this kind of upstream coding region, that at some point overlaps the emergent start-codon.
This sequence, then, would be interrupted by the start-codon created by the frameshift mutation. Additionally, overlapping the newly created reading frame, should be the remaining coding sequence, now shifted by a single nucleotide forward. So the ATG start codon for NylB should have two single-nucleotide overlaps with two codons from the previous sequence.
The A in the ATG start-codon for NylB should be overlapped by the last codon in the upstream sequence, and the G in the ATG start codon should be overlapped by the first codon in the remainder of the original reading frame.
This is exactly what we see. I conclude there has been no gaff.
There is none.
I only used the amino acid sequence you provided, (which does appear to terminate prematurely, yes) but I’ll tell you why, because I looked at Okata H et al 1983 and saw the original nucleotides sequences for P-NylB and F-NylB.
I recommend you do this now so you can see what I mean. First I looked at the upstream region before the coding region for NylB begins, and saw that despite this region not coding for the NylB sequence, it was just as highly conserved as the NylB coding region. This is what originally led me to believe that sufficient quantities of Ohno’s hypothetical ancestral reading frame still might contain enough information to produce significant alignments.
Then I looked through the coding region for NylB and saw that the conservation of sequence pretty much continued there.
But then when we get to the end of NylB and go beyond the NylB stop-codon, this level of conservation very clearly drops off immediately following the NylB stop codon. As you can see, the nucleotide sequences from the F-NylB and P-NylB plasmids diverge pretty much completely. This tells me there’s no more information about the original protein left to bother searching for it.
Which indicates that even though there hypothetically might have been even more protein coding sequence in that region, it would be a waste of time trying to get any meaningful alignments for that part of the protein.
So while I could seach for alignments for that portion, I don’t think they would be significant. As in, even should I happen to find something that aligns to this portion, it is small enough, and so affected by lack of conservation, I wouldn’t believe the result I got. In effect, I would have to decide whether I thought any of the two plasmids had faithfully retained the ancestral stage for this region. And I simply don’t when they so clearly diverge following the NylB stop codon.
Heh, funny stuff. 🙂
Regardless, I’ll post the results of the structural prediction and alignment search I got in my next post.
No, that would be my view too.
Looks to me like Sal is trying to dismiss the fact that you found that the embedded relic sequence does in fact fold, because, you know, evolutionists claim that most random sequences will fold anyway!!1!!1
Anyway, so I used the amino acid sequence Sal was persistent enough to bother compiling for us, back here.
I assumed Sal had been thorough so I went on to do some NCBI protein BLAST searches myself. The first thing I found was an amino acid alignment for this beginning (aka n-terminal) portion of the protein:
MGYIDLSAPVAMIVSGGLYYLFTRRGYTFGDTR
Which produced a ~97% identical alignment for the c-terminal (aka the end) of the nitrate reductase enzyme from Leucobacter celer.
Another stretch of Ohno’s hypothetical protein (the part that overlaps the whole NylB coding region), an 247 amino acid stretch at the c-terminal end, aligns to an uncharacterized protein from the fungus Pneumocystis murina (strain B123), with a sequence-similarity of about 29%.
These findings, combined with what I described in this post above led me to believe it really might be worth it to submit the amino acid sequence for Ohno’s PR.C sequence to a protein structural prediction and alignment service of some sort.
So that’s what I did. I went to the I-TASSER server for protein structure and function prediction, and submitted the nucleotide sequence. And then I got this in the mail this morning: http://zhanglab.ccmb.med.umich.edu/I-TASSER/output/S341628/
Now I’m no expert, so how exactly this server computes how the amino acid sequence will fold is beyond me. There are some nice little helpful descriptions though. I scrolled down to the part titled “Top 5 final models predicted by I-TASSER” and looked at them. The “Top” predicted model, the server informs, is used for alignment to known protein structures.
The best match in alignments is shown in the “Proteins structurally close to the target in the PDB (as identified by TM-align)” section.
As this section informs, “(b) Ranking of proteins is based on TM-score of the structural alignment between the query structure and known structures in the PDB library.”
The protein I submitted got a top TM-score of 0.914. As the website informs “TM-score is a metric for measuring the structural similarity of two protein models. It is designed to solve two major problems in the traditional metrics such as root-mean-square deviation (RMSD): (1) TM-score measures the global fold similarity and is less sensitive to the local structural variations; (2) magnitude of TM-score for random structure pairs is length-independent. TM-score has the value in (0,1], where 1 indicates a perfect match between two structures. Following strict statistics of structures in the PDB, scores below 0.17 corresponds to randomly chosen unrelated proteins whereas with a score higher than 0.5 assume generally the same fold in SCOP/CATH.“
So with my non-expert understanding, I take this to mean that since the best matching protein I got a score of 0.914 with a coverage of 93% of the full sequence, his is a very significant match to a real protein with a known function.
What is the most intuitively accesible thing is that there is a nice spinnable 3D model of the structural alignment, so you can really see how well it matches. The sequence used is shown in rainbow colors, and the protein it aligns to is shown as a purple string. I have attached a picture, but unfortunately it’s really hard to capture how well it matches in a single small pic for a large 3dimensional structure. Go to the results page and look at the 3d model and spin it around, zoom in and out and so on.
Interestingly, very nearly the whole protein aligns, but to a significantly larger protein. Whether this indicates that Ohno’s inferred ancestral protein is a throwback to a simpler version, or just that what we see is the remnants of something that used to be bigger but has been erased to time is anyone’s guess right now.
If you click on the “PDB hit” link provided, you will see that the ancestral protein actually adopts the same fold as the eukaryotic (in this case, human) nuclear pore complex protein. From there you can click on structure similarity and see there are many many other proteins that adopt this conformation.
By the way, that is all of Ohno’s inferred ancestral sequence that aligns. Both the portion before (as in upstream and not overlapping the NylB coding region), and the part after that fully overlaps the NylB coding region. This strongly implies the belong to the same protein, and as such that NylB really was created by a frameshift mutation 33 codons downstream in that ancestral gene.
I should add that the blue region is the n-terminal region (and thus the part that is upstream to the NylB region). This is the part that forms beta sheets and seem to be the portion that creates some sort of membrane pore/channel (or anchor point) in the larger protein.
Here’s a model of the nuclear pore complex. Ohno’s ancestral protein would be homologous to the orange ones in this picture.
Correction, it should read “.. for NylB, but only by a single nucleotide“. In case there is any confusion.
that’s pretty cool. But we’ll need to compare the results to the ID’s software counterpart for good measure
function looks_designed(string seq)
return true;
Rumraket,
Very cool. When you say homologous, how much sequence similarity do you mean?
Yeah and what you found aligned with the UNSHIFTED portion! Not exactly vindication of Ohno. This like the sixth time you been informed of your error, and you still keep trumpeting it. Pathetic.
Really cool Rumraket, you showed how Ohno’s typo created a fold. Not exactly reassuring to your thesis, unless you want to claim how miraculous it was that his typo accidentally helped you create a new hypothesis.
N
I’ve refuted this now three times at least you silly and deceptive misinformer.
Here is Ohno’s hypothetical protein:
MGYIDLSAPVAMIVSGGLYYLFTRRGYTFGDTRERTFHRPAPRQVSRSRGRGADTRQLAGGPAQPLGLRPPGRAAAHGGGLPARPGDARGARRAARRARDAAPRSRAAARGDLHRRIPRAARLRGPRRVLPGGFRTRRPSPADERLEVAVRHGRRRADRRGAHRSRAARHRVCTRARGLRLRRALRAAGARHADLDRLQRGLRRSGLGGADPRSLRRLAHAARRGPRRHLRVPHHPPRRRRHRRVPVLLGEHRRARLDRRAGHRSALRRSALHVPVGEARRRSGCDHHGRPDRLRLRERGRLLHRAGSRTRGPHDARRRRRSRRTGRIAGLGGKRAGRRLPRSHDRRGFHLRIPRGQLHAPVVVHGQRARQRERHRHPRPEPLARSAHRLRDRQALVVARSRHPALARAAERDPARRQPCPRRGVGG
Here is the best closes sequence Rumraket’s theatrics could generate. It was ranked number 1. What would be persuasive of a frame shift is that the number 1 ranking protein would at least look similar sequence wise. It doesn’t!
QYQEALSHPLALHLNTNMVCLLKKGYLSFSSLVDHLYLLPYENLLTEDETDDVDIRDVICLIKCLRLVTVDMSVIMEMSCYNLQSPEKAAEQILEDMITIDVENVMEDICSKLQENPIHAIGLLIREMDYETRMNLTQLYGSNTAGYIVCRGVHKIASTRFLICRDLLILQQLLMLLSYYLIKWGSECLATDNRFVIVELFFQEVAHILSQTGLNWPEMITAITSYLLQLLWPSNPGCLFLENCQYVQLQDFMLGRCYLVTGEGQKALECFCQAFLDRLIRSEDGEIVSTPRLQYYDKVLRLLDVIGLEAGDDWKSQATLRTCIFKHHLDLGHNS-----------------------------QAYEALTQIPDSSRQLDCLRQLVVVLCERQDLVEFPYVNNEVVGIIESRARAVDLMTHNYYEL
Echoing colewd, I thank Rumraket for engaging this discussion, but a more convincing evidence of Ohno’s frame shift happening millions of years ago would look more like:
MGYIDLSAPVAMIVSGGLYYLFTRRGYTFGDTRERTFHRPAPRQVSRSRGRGADTRQLAGGPAQPLGLRPPGRAAAHGGGLPARPGDARGARRAARRARDAAPRSRAAARGDLHRRIPRAARLRGPRRVLPGGFRTRRPSPADERLEVAVRHGRRRADRRGAHRSRAARHRVCTRARGLRLRRALRAAGARHADLDRLQRGLRRSGLGGADPRSLRRLAHAARRGPRRHLRVPHHPPRRRRHRRVPVLLGEHRRARLDRRAGHRSALRRSALHVPVGEARRRSGCDHHGRPDRLRLRERGRLLHRAGSRTRGPHDARRRRRSRRTGRIAGLGGKRAGRRLPRSHDRRGFHLRIPRGQLHAPVVVHGQRARQRERHRHPRPEPLARSAHRLRDRQALVVARSRHPALARAAERDPARRQPCPRRGVGG
NOT
QYQEALSHPLALHLNTNMVCLLKKGYLSFSSLVDHLYLLPYENLLTEDETDDVDIRDVICLIKCLRLVTVDMSVIMEMSCYNLQSPEKAAEQILEDMITIDVENVMEDICSKLQENPIHAIGLLIREMDYETRMNLTQLYGSNTAGYIVCRGVHKIASTRFLICRDLLILQQLLMLLSYYLIKWGSECLATDNRFVIVELFFQEVAHILSQTGLNWPEMITAITSYLLQLLWPSNPGCLFLENCQYVQLQDFMLGRCYLVTGEGQKALECFCQAFLDRLIRSEDGEIVSTPRLQYYDKVLRLLDVIGLEAGDDWKSQATLRTCIFKHHLDLGHNS-----------------------------QAYEALTQIPDSSRQLDCLRQLVVVLCERQDLVEFPYVNNEVVGIIESRARAVDLMTHNYYEL
But Rumraket managed to prove a sequence generated through a deliberate deletion and an accidental typo can converge on similar fold. He could try correcting Ohno’s typo and seeing what fold that will converge on! 🙂
That shows exactly why I have reluctance accepting fold improbabilities that other creationists do. I’ve gone on record saying so.