On Deciphering the Signature

Bookmark and Share

September 12, 2011 Tags: Design

Today's entry was written by Darrel Falk. You can read more about what we believe here.

On Deciphering the Signature

Steve Meyer has responded to Dennis Venema’s review1 of his book Signature in the Cell in the September 2011 issue of Perspectives on Science and the Christian Faith (PSCF) (63:171-182). Although, Dennis has ably responded (63:183-192), I would like to address one specific aspect of Meyer’s response, especially since it relates to the final paragraph of my initial essay regarding the book and Dennis’s six part series on the BioLogos website.

BioLogos has dealt fairly extensively with what we thought was the basic premise of Signature in the Cell. I had read the book carefully and I know Dennis did as well before we responded. I sincerely thought that the heart of Meyer’s argument is summarized in the following three quotes from the book:

1. “So the discovery of the specified digital information in the DNA molecule provides strong grounds for inferring that intelligence played a role in the origin of DNA. Indeed, whenever we find specified information and we know the causal story of how that information arose, we always find that it arose from an intelligent source. It follows that the best, most causally adequate explanation for the origin of the specified, digitally encoded information in DNA is that it too had an intelligent source. Intelligent design best explains the DNA enigma” (p. 347, emphasis added).

2. “Since, as argued in Chapters 8 through 15, intelligence is the only known cause of large amounts of specified information, the presence of such information in the cell points decisively back to the action of a designing intelligence” (p. 382, emphasis added).

3. “Because we know intelligent agents can (and do) produce complex and functionally specified sequences of symbols and arrangements of matter, intelligent agency qualifies as an adequate causal explanation for the origin of this effect. Since, in addition, materialistic theories have proven universally inadequate for explaining the origin of such information, intelligent design now stands as the only entity with the causal power known to produce this feature of living systems.” (p. 386, emphasis added).

So we at BioLogos have always thought that if mainstream science demonstrated an increase in “complex specified information” (CSI) without needing to invoke supernatural intervention, Meyer’s assertion that “intelligence is the only known source of such information in the cell” will have been refuted at the scientific level. It sure seemed to me that this is what he said in the above quotes.

With that in mind, we’ve put a great deal of effort into showing a number of cases in the lab and in nature where scientific data have provided very strong evidence for increased CSI which is entirely consistent with how we scientists would define “natural explanations.” All this time, starting with my first essay almost two years ago, we sincerely thought we were engaging Meyer’s book on Meyer’s terms.

But now, in his PSCF article, Meyer states that arguments based on examples of increased CSI don’t count if they occur after life began on Earth.

Signature in the Cell argues, first that no purely undirected physical or chemical process—whether those based upon chance, law-like necessity, or the combination of the two—has provided an adequate causal explanation for the ultimate origin of the functionally specified biological information. In making that claim, I specifically stipulate that I am talking about undirected physical and chemical processes, not processes (such as random genetic mutation and natural selection) that commence only once life has begun. Clearly material processes that only commence once life has begun cannot be invoked to explain the origin of information necessary to produce life in the first place) (pp. 173-174, Perspectives on Science and Christian Faith, Sept. 2011, emphasis added).

Since I had read the book very carefully, and have gone over it many times since, I was amazed that I could have missed this stipulation. Again, he says: “I specifically stipulate that I am [not] talking about … processes (such as random genetic mutation and natural selection) that commence only once life has begun.”

Did he really specifically stipulate that? Have we been barking up the wrong tree all this time? While we knew the main focus of Meyer's book was the origin of life (not mechanisms of evolution), his argument clearly stated, we thought, that no large increase in CSI (Complex Specified Information) had ever been demonstrated without the need to invoke intelligence. Period.

I went back through my well-marked up copy of the book again, re-examining each section in which he wrote about increased CSI. Despite my best efforts, I could not find the stipulation he mentions in the PSCF article. Still, thinking I had missed it, I spent $15 for an electronic version of the book—one that would allow me to identify every time the word “mutation,” or natural selection” appeared—anything that would help me find his stipulation. I couldn’t find it.

Actually I thought Meyer was pretty clear and highly specific in his book. Consider this scientific challenge on page 429:

If, for example, someone successfully demonstrated that "large amounts of functionally specified information do arise from purely chemical and physical antecedents," then my design hypothesis, with its strong claim to be the best (clearly superior) explanation of such phenomena, would fail.

Find a case where a large amount of CSI has accumulated without needing to invoke intelligence, and his argument, Meyer said, fails. This is a strong statement, clearly worded, and there is no hint of Meyer’s stipulation that it doesn’t count if life has already begun. In Dennis Venema's BioLogos blog series, he showed many cases where there were large increases in CSI (whole genome duplication, for example) without needing to invoke that supernatural intervention was necessary to create it. Chromosomes, the cell division machinery, and nucleotides are “purely chemical and physical antecedents.” The information content in the genome, Venema showed, quadrupled early in vertebrate history through material processes that we know and understand well. Did this not meet the scientific criteria that Meyer specifically called for?

I don’t know how misunderstandings like this happen. I believe that Stephen Meyer, who I consider to be a friend and colleague, thinks the stipulation exists in his book and that he worded it clearly. I assume he thinks it was implied in some overarching statement that I have not been able to find. I also think he believes he was clear. Unfortunately, clear he was not. I’ve looked thoroughly and I have not been able to find his stipulation.

In post after post, we have set out to demonstrate the scientific case we thought Meyer called for. Then in the end, it sure seems to us, that the rules changed, even though Steve feels they were written in his book all the way along.

Still, let’s move on. Let’s play by the new rule and let’s define it carefully.

So here’s the rule as I now understand it: If large increases in CSI can be demonstrated without the need to invoke an external intelligence, “then [Meyer’s] design hypothesis with its strong claim to be the best (clearly superior) explanation of such phenomena, would fail.”

Having stated the rule, we have to make two exceptions (Meyer himself made Exception #1 clear in Chapter 13; Exception #2 is the new stipulation we've been discussing):

Exception 1. We can’t count large increases in CSI which develop as a result of computer programs because minds designthe program parameters.

Exception 2. We can’t count large increases in CSI which develop in the history of life, because DNA was necessary to set those processes in motion.

So what can we count? Until he clarified the existence of Exception #2, I thought any general increases in CSI would count. However, it is now very hard for me to imagine any increase in information that would not be categorized within either Exception 1 or Exception 22. The only thing left that doesn’t fit into one of these two exceptions is the origin of life itself. The point of the book, I thought, was to bring other examples of increased CSI to bear on this very question.

With Meyer’s exceptions and the inability to bring general CSI increases to bear on the origin of life question, we also no longer have “positive3 experiments [which] provide causal adequacy of intelligent design” (p. 335, emphasis added).

So what are we left with? Are we not simply left with the question of whether the origin of life experiments show that information-rich molecules will arise in a test tube from chemicals off the shelf? Dr. Meyer, I think, says no, for reasons that are no longer clear to me other than that he’s given up on the science. I, on the other hand say, “Wait a while. Let the science play itself out before a scientifically based decision is made.” To be frank though, I am a little concerned that even if the right mix of materials is found to produce molecules that can spontaneously assemble in a manner that gives rise to complex specified information, Dr. Meyer or those who follow him will say, “Sorry, you can’t count that because it took a mind to create the conditions and it took a mind to mix them together in a test tube.” And with that we’ll have a new stipulation which most likely was in some manner implied in Signature in the Cell to begin with.4

The interesting thing about this is that Steve Meyer and I are probably really in almost the same exact position when it comes to our core beliefs. Obviously as fellow Christians, we both believe that there is a Mind behind the process. We both think that the history of life with its constant increase in complex specified information is a product of the activity of God. We both stand amazed at the majesty of creation and our love for the Creator who is personally involved not only in our own individual lives but those of our families and faith communities as well. We differ primarily in one regard. Steve thinks he has shown through scientific analysis that this Mind we both believe in must have been present and supernaturally active in the creation of information. I think the Mind (God) was present, but I can’t put the existence of God into a scientific experiment to demonstrate God's activity. Furthermore, unlike Steve, I have no pre-conceived ideas about whether God's,supernatural activity was necessary for creation of information. God, as I see it, may have chosen to create information bearing molecules indirectly through God’s natural activity in a manner that is analogous to the development of a baby or the growth of a tree from a seed.

In the end, our difference is simple, he thinks that the test tubes won’t ever deliver information rich molecules and I think it is too early to say. He has declared the matter more or less settled on the basis of scientific analysis. I consider the matter fully unsettled. But the most important thing of all has been settled and on this we both agree. This Mind we speak of is God’s Mind--God's Holy Spirit. That Spirit not only fills all of creation, but more specifically that Spirit fills us with his Presence and envelopes us in his love. This is cause for celebration and, with "sandals off," we each bow our heads in humble worship. Truly, we--all of us--are standing on holy ground.

Notes

1. Perspectives in Science and Christian Faith 62:276
2. Note to Steve: Does not the human brain count within Exception #2? After all, it arose in the history of life and its development depends upon DNA. If so, you might need an exception to the exception.
3. The term “positive” is used 21 times in the book. It is clearly important to the author that the evidence for intelligence associated with the origin of DNA be viewed not as absence of contrary evidence, but rather a piece of convincingly positive evidence that hinges upon the fact that CSI in general, can’t be built without a mind.
4. I’m really not trying to be facetious here. I really do think that’s what would happen. I can almost draft the stipulation now.


Darrel Falk is former president of The BioLogos Foundation. He transitioned into Christian higher education 25 years ago and has given numerous talks about the relationship between science and faith at many universities and seminaries. He is the author of Coming to Peace with Science.


View the archived discussion of this post

This article is now closed for new comments. The archived comments are shown below.

Loading...
Page 3 of 3   « 1 2 3
Genomicus - #64901

September 21st 2011

John: (continued from above)

Now, this
doesn’t mean that we should despair and explaining how the sequence similarity
shared between these six proteins arose. We should not throw our hands up in
the air and declare that “God did it.” Instead, I propose an explanation for
this sequence similarity that does not entail common descent. I propose that
the shared similarities can be explained by functional features instead of an
evolutionary origin.

 

Indeed, as Mike Gene pointed out in 2003, researchers expected these flagellar components to share sequence similarities because of their functional properties:

 

“Some degree of similarity among the sequences of the axial proteins would not be surprising, for two reasons. The first is that, since the axial proteins together form a continuous filamentous structure, we might expect them to have a similar lattice and to share common structural elements that determine the lattice….The second reason for looking for sequence similarities among the axial components concerns the manner in which they are thought to be exported across the cell membrane.” (Homma, M, DeRosier, DJ, and Macnab, RM. 1990. Flagellar hook and hook-associated proteins of Salmonella typhimurium and their relationship to other axial components of the flagellum. J. Mol. Biol. 213: 819-832)

 

 

Thus, the response to your question “how do you explain” the homology shared between the axial proteins is that these proteins are homologous because of their functional properties. In other words, because of similar function, subcellular location, similar constraints, etc., we would expect convergent evolution at the molecular level to occur among these six proteins.

 

I will even go so far as to propose a prediction for this hypothesis:

Consider that, for example, 41 residues in FlgG and
FlgC are identical. If this level of similarity was the result of convergent
evolution, then I would expect that only about 1/3 of the codons that code for
these 41 residues would be the same. Changes in synonymous codons are far more
neutral than changes at the amino acid level, so if these 41 identical
positions happened to come about independently, via convergent evolution, then
chance dictates that about 1/3 of the codons that code for these 41 positions
would be identical (assuming that, on average, there are about 3 codons that
code for one amino acid). Given that, in bacteria especially, there actually is
some selective pressure on synonymous mutations, we would expect that the
number of identical codons (out of the 41 codons) between the two proteins to
be just a little above 33%, probably anywhere between 33% - 40%. This is what
my hypothesis suggests, and if confirmed, this would certainly bolster the
suggestion that the sequence similarities between the axial proteins is the
result of convergent evolution, instead of a common origin. What do you predict
John? Do you think that more than 33% - 40% of the 41 codons will be identical?
What does the Darwinian hypothesis predict?
<input id=“gwProxy” type=“hidden”><!—Session data—><input id=“jsProxy” type=“hidden”><div id=“refHTML”>


John - #64921

September 22nd 2011

“I will even go so far as to propose a prediction for this hypothesis:”

br>

It’s not whether YOU make a prediction, but whether the hypothesis does independent of you or me. If we can’t agree on the predictions of a hypothesis, it’s not a good hypothesis.

br>

“Consider that, for example, 41 residues in FlgG andbr> FlgC are identical. If this level of similarity was the result of convergent evolution, then I would expect that only about 1/3 of the codons that code for these 41 residues would be the same.”

br>

That’s nice, but you are ignorant of codon bias.

br>

“Changes in synonymous codons are far more neutral than changes at the amino acid level, so if these 41 identicalbr> positions happened to come about independently, via convergent evolution, then chance dictates that about 1/3 of the codons that code for these 41 positions would be identical (assuming that, on average, there are about 3 codons that code for one amino acid).

br>

There aren’t.


John - #64922

September 22nd 2011

“...we would expect that the number of identical codons (out of the 41 codons) between the two proteins to be just a little above 33%, probably anywhere between 33% - 40%. This is what my hypothesis suggests,”

br>

Only if codon bias doesn’t exist.

br>

” and if confirmed, this would certainly bolster thebr> suggestion that the sequence similarities between the axial proteins is the result of convergent evolution, instead of a common origin. What do you predict John?”

br>

The issue is not what I personally predict. I doubt that you can get that basic concept through your head.

 

“Do you think that more than 33% - 40% of the 41 codons will be identical?”

br>

Yes, but it has nothing to do with convergence or lack thereof. 

br>

You might want to look into mutation rates too, even if you only accept a 6000-year-old earth.


Genomicus - #64903

September 21st 2011

Ashe:

>> Do you have a reference? <<

Yes:

”...structural information in a membrane protein
sequence can be statistically interpreted.
Elements of the structural simplicity of these proteins
suggest the existence of commonly used patterns
in transmembrane helix-helix interactions.
First, the space that natural selection can sample in
search of favorable combinations seems to be limited
by the low complexity of the sequences, since
two-thirds of transmembrane residues comprise,
on average, only six amino acids (Leu, Ile, Val,
Phe, Ala, and Gly)....”

(Alessandro Senes, Mark Gerstein and Donald M. Engelman. Statistical Analysis of Amino Acid Patterns in Transmembrane Helices: The GxxxG Motif Occurs Frequently and in Association with b-branched Residues at Neighboring Positions. J. Mol. Biol. (2000) 296, 921-936)

That’s why Cascales et al. states that, with regards to the homology shared by MotA/B and the Tol-Pal system (which TolQ/R are components of):
“...the relevance of these sequence alignments is difficult to quantify, depending on a
large alignment of a relatively large number of proteins from a wide
selection of organisms distributed in a rather non-random fashion across
the phylogenetic space.”

(E. Cascales, R. Lloubès, and J. N. Sturgis, “The TolQ–TolR proteins
energize TolA and share homologies with the flagellar motor proteins
MotA–MotB,” Molecular Microbiology, Vol. 42 (3):795-807)<input id=“gwProxy” type=“hidden”><!—Session data—><input id=“jsProxy” type=“hidden”><div id=“refHTML”>


Ashe - #64912

September 22nd 2011

From Cascales et. al. 
Interestingly, in the MotA sequences, this region is less conserved, despite its presumed functional importance in the formation of an ion pore. Comparing the N-terminal regions of the TolR, ExbD and MotB proteins gives strong consensus sequences and, as in TolQ hairpin, it is possible to observe a homology between the different groups of sequences, in particular the aspartate present in the overall 90% consensus sequence (Fig. 4C). This residue has previously been found to play a major role in both TonB function (Braun et al., 1996) and cell motility (Zhou et al., 1998b).”
br>

pre>In NCBI conserved domain database these three proteins are together. actually
most annotations for homologous sequences without experimental data are like
MotA_ExbB, or MotA/TolQ/ExbB proton channel family.  please see the following url
for NCBI CD entry for MotA/TolQ/ExbB proton channel family
http://www.ncbi.nlm.nih.gov/Structure/cdd/cddsrv.cgi?uid=186086

br>
br>

Genomicus - #64904

September 21st 2011

Ashe:

You are correct, of course, when you say that FliI shares sequence similarity with the beta subunit of the ATP synthase. I glossed over that homology. FliI does indeed have a homologue that is non-flagellar.
<input id=“gwProxy” type=“hidden”><!—Session data—><input id=“jsProxy” type=“hidden”><div id=“refHTML”>
WhiteSwan - #64905

September 21st 2011

I take it you’re not impressed by the observation that energy transduction through the TolA and TonB proteins  is mediated by protein homologs to MotAB. Besides its requirement for cell integrity and for LPS expression to the cell surface, afaik, we do not know the precise function of the Tol system. 


Ashe - #64906

September 21st 2011

Whoops logged in with an older username 


beaglelady - #64916

September 22nd 2011

Genomicus, are you the same person as Kirk Durston?


Genomicus - #64924

September 22nd 2011

John:
<!—[if gte mso 9]><xml>

Normal0

</xml><![endif]—>

>> It’s not whether YOU make a prediction, but whether the hypothesis does independent of you or me. If we can’t agree on the predictions of a hypothesis, it’s not a good hypothesis. <<

 

Yea, but the hypothesis I outlined to explain the observed sequence similarity between FlgBCEFGK does predict that only about 33% - 40% of the codons encoding the identical residues in the proteins would themselves be identical.

 

>> That’s nice, but you are ignorant of codon bias. <<

No, I’m not. I took codon-usage bias into account when making my prediction. That’s why I said:

“Given that, *in bacteria especially, there actually is
some selective pressure on synonymous mutations*, we would expect that the
number of identical codons (out of the 41 codons) between the two proteins to
be just a little above 33%, probably anywhere between 33% - 40%.”

 

I am well aware of codon-usage bias, John. That’s why we wouldn’t predict exactly 33% of the codons to be identical. We would predict a few percent more than 33% of the codons to be identical, precisely because of codon-usage bias.

 

>> There aren’t [3 codons on average that code for one amino acid]. <<

 

Yea but Dr. Douglas Theobald says that (in his web article, “29+ Evidences for Macroevolution”):

“The genetic code itself is informationally redundant; on average there are three different codons (a codon is a triplet of DNA bases) that can specify the exact same amino acid….”

Why don’t you send him an email to correct his ‘error’?

 

His source is Voet and Voet, 1995 (Voet, D., and Voet, J. (1995) Biochemistry. New York, John Wiley and Sons). With all due respect, I’d prefer to follow the statement of a credible source rather than online person named John.

 

>> You might want to look into mutation rates too, even if you only accept a 6000-year-old earth. <<

 

I don’t accept a 6000-year-old earth anymore than you do. Why did you choose the number 6000 instead of 23,450, or 58,000? Or is it perhaps that you assume (or partially assume) that every teleologist will accept a 6000 year old earth?

 

Having said that, I’d like to emphasize the following point:

That the hypothesis that FlgBCEFGK share their
sequence similarity because of convergence does indeed predict that out of the
total number of codons encoding identical amino acids, only 33% - 40% of these
codons will be identical. The reason we extend the number 33% to 40% is
precisely because of codon-usage bias. On the other hand, John, you expect that
the percent of identical codons will be more than 40%. Why is that?
Furthermore, please point out exactly why you disagree with my suggestion that
the convergence hypothesis predicts that specific outcome?

<input id=“gwProxy” type=“hidden”><!—Session data—><input id=“jsProxy” type=“hidden”><div id=“refHTML”>


John - #64930

September 23rd 2011

span style=“margin-top: 0px; margin-right: 0px; margin-bottom: 0px; margin-left: 0px; padding-top: 0px; padding-right: 0px; padding-bottom: 0px; padding-left: 0px; border-top-width: 0px; border-right-width: 0px; border-bottom-width: 0px; border-left-width: 0px; border-style: initial; border-color: initial; vertical-align: baseline; “>

Me: >> That’s nice, but you are ignorant of codon bias. <<br>
“No, I’m not. I took codon-usage bias into account when making my prediction. That’s why I said:

“Given that, *in bacteria especially, there actually isbr> some selective pressure on synonymous mutations*, we would expect that the number of identical codons (out of the 41 codons) between the two proteins to be just a little above 33%, probably anywhere between 33% - 40%.””

br>

And that is why *I* said that you are ignorant of codon bias. But what’s amazing is that you lack the courage and/or mental horsepower to simply do the alignment!

 

“I am well aware of codon-usage bias, John. That’s why we wouldn’t predict exactly 33% of the codons to be identical. We would predict a few percent more than 33% of the codons to be identical, precisely because of codon-usage bias.”

br>

And why would, uh, “we” predict that only a few percent more would be identical?

 

>> There aren’t [3 codons on average that code for one amino acid]. <<

 

“Yea but Dr. Douglas Theobald says that (in his web article, “29+ Evidences for Macroevolution”):”

br>

Hilarious! This shows how intellectually crippled you are. I wrote something that was plainly false, but instead of answering back, “61 codons / 20 aa residues = ~3,” you quote someone. 

br>

Do you have any idea how insane that is?

br>

>> You might want to look into mutation rates too, even if you only accept a 6000-year-old earth. <<

 

“I don’t accept a 6000-year-old earth anymore than you do. Why did you choose the number 6000 instead of 23,450, or 58,000? Or is it perhaps that you assume (or partially assume) that every teleologist will accept a 6000 year old earth?”

br>

Why aren’t you looking into mutation rates? The age of the earth was merely an aside.

 

“Having said that, I’d like to emphasize the following point:

That the hypothesis that FlgBCEFGK share theirbr> sequence similarity because of convergence does indeed predict that out of the total number of codons encoding identical amino acids, only 33% - 40% of these codons will be identical.”

br>

No, the hypothesis, in the context of what we know about genetics, does not.

br>

“The reason we extend the number 33% to 40% isbr> precisely because of codon-usage bias.”

br>

And how did you calculate that?

br>

“On the other hand, John, you expect thatbr> the percent of identical codons will be more than 40%. Why is that?”

br>

Because of codon bias and mutation rates. You don’t understand the former and you ignore the latter.

span style=“font: 18.0px ‘Lucida Grande’”>
“Furthermore, please point out exactly why you disagree with my suggestion thatbr> the convergence hypothesis predicts that specific outcome?”

br>

Because of codon bias and mutation rates. You don’t understand the former and you ignore the latter.

br>

Again, your need to personalize everything is fatally crippling your ability to think critically.

/p>


Genomicus - #64925

September 22nd 2011

Ashe:

>>I take it you’re not impressed by the observation that energy transduction through the TolA and TonB proteins  is mediated by protein homologs to MotAB.<<

It depends on the level of sequence similarity. If the homology is established on the grounds of structural similarity primarily, then I wouldn’t be all that impressed. What proteins are you referring to?


<input id=“gwProxy” type=“hidden”><!—Session data—><input id=“jsProxy” type=“hidden”><div id=“refHTML”>
Genomicus - #64926

September 22nd 2011

beaglelady:

No, I’m not.
<input id=“gwProxy” type=“hidden”><!—Session data—><input id=“jsProxy” type=“hidden”><div id=“refHTML”>
Genomicus - #64943

September 23rd 2011

John:

<!—[if gte mso 9]><xml>

Normal0

</xml><![endif]—>

>> And why would, uh, “we” predict that only a few percent more would be identical?<<

Precisely because of codon-usage bias. The rest of your response, unfortunately, consists of little more than calling me (or my comments) “hilarious” or “ignorant” or “insane.” So I will respond to one particular thing you said, and something that I think is most revealing.

You said,

“…what’s amazing is that you lack the courage and/or mental horsepower to simply do the alignment!”

I actually did this last night to check the prediction of the convergent-evolution hypothesis. The results are below.

FlgG and FlgC Analysis

Using ClustalW, I aligned an FlgG (accession number P0ABX5) sequence from Escherichia coli and an FlgC (accession number P0ABX2) sequence from E. coli. A total of 41 identical residues were shared between the two proteins. I then retrieved the DNA sequences of those two proteins from GenBank, and using Pal2Nal, I determined which codons encoded the identical residues. Only 40 codons were, however, used for the analysis, since one of the 41 codons was methionine, and only one codon specifies methionine.

The results are as follows:

Out of 40 codons, 15 of them were identical. Thus: 15/40 = 37.5% of the total codons are identical, consistent with the prediction that 33% - 40% of the codons will be identical.

FlgB and FlgE Analysis

ClustalW was used to align an FlgB (P0ABW9) protein sequence from Escherichia coli and an FlgE (P75937) protein sequence from E. coli. A total of 22 identical residues were shared between the two proteins. I then retrieved the DNA sequences of those two proteins from GenBank, and using Pal2Nal, I determined which codons encoded the identical residues. Only 21 codons were, however, used for the analysis, since one of the 22 codons was methionine, and only one codon specifies methionine.

The results are as follows:

Out of 21 codons, 8 of them were identical. Thus: 8/21 = ~38% of the total codons are identical, consistent with the prediction that 33% - 40% of the codons will be identical.

This is just a cursory analysis, investigating 2 protein combinations out of a total of about 30 different combinations. I suspect that further analysis will reveal that, in general – with a few exceptions here and there – 33% - 40% of the codons will be identical.

I’ll be looking into more sequences this weekend,
during which time I won’t be responding to any of you. I’ll respond on Monday,
hopefully. In the meanwhile, the above facts are available for you to digest,
John: namely, the fact that, so far, my prediction does seem to be faring
pretty well, and of course it refutes your hypothesis that I lack the “the
courage and/or mental horsepower to simply do the alignment.”

<input id=“gwProxy” type=“hidden”><!—Session data—><input id=“jsProxy” type=“hidden”><div id=“refHTML”>


John - #64945

September 23rd 2011

Genomicus:

“The rest of your response, unfortunately, consists of little more than calling me (or my comments) “hilarious” or “ignorant” or “insane.” “

That’s simply false and you know it. Let me copy the parts that you are refusing to address:

“Why aren’t you looking into mutation rates? The age of the earth was merely an aside.”

“And how did you calculate that [codon usage bias]?”


These two factors have a much larger magnitude than anything predicted by your hypothesis. Your refusal to address them demonstrates that at some level, you understand this.

Genomicus - #65060

September 26th 2011

John:

>> “These two factors have a much larger magnitude than anything predicted
by your hypothesis. Your refusal to address them demonstrates that at
some level, you understand this.” <<

Yea, but the prediction made by the convergent evolution hypothesis (CEH, for short) to explain the sequence similarity shared among FlgBCEFGK is based upon biological observation, instead of being an argument “on paper,” like your argument is. If we take a random protein and align it with another protein, and use the methodology outlined above in my previous response, we will see that anywhere in between 33% - 40% of the codons that encode identical residues will themselves be identical.

An actual biological example may be used here:
No homology has been suggested for YscJ and YscT. They are different proteins that have different ancestries. If one aligns YscJ and YscT (using ClustalW) there will be 29 identical positions. If one threads the amino acid alignment into a codon alignment (using Pal2Nal), one finds that, out of 29 total codons that encode identical residues, only about 10 of those codons are themselves identical.

Thus: 10/29 = ~34.5% of the total
codons are identical, a number that lies within the 33% - 40% range I described. It looks like codon-usage bias and mutation rates didn’t affect the percentage of identical codons as drastically as you envision.

Your arguments might look fine on paper, but I’m basing my arguments on biological reality. You can argue all you want regarding codon-usage bias and mutation rates, but simple experiments in bioinformatics like the one described above answer your arguments.



<input id=“gwProxy” type=“hidden”><!—Session data—><input id=“jsProxy” type=“hidden”><div id=“refHTML”>


John - #65069

September 27th 2011

”…instead of being an argument “on paper,” like your argument is.”

My argument is based on actual bioinformatics that I’ve used in the lab for more than a decade.

“If we take a random protein and align it with another protein, and use the methodology outlined above in my previous response, we will see that anywhere in between 33% - 40% of the codons that encode identical residues will themselves be identical.”

No, we won’t, because if we tabulate ALL the coding sequences for E. coli, we will see that the usage of a particular codon (for multiple codons encoding the same residue) varies between 

You’re not being honest at all, because you didn’t take a random protein. You tried one and when it gave you the answer you desired, you quit.

“An actual biological example may be used here:…”

No need for an example, Kirk. We have THE WHOLE GENOME.

”…If one threads the amino acid alignment into a codon alignment (using Pal2Nal), one finds that, out of 29 total codons that encode identical residues, only about 10 of those codons are themselves identical.”

And if one takes THE WHOLE GENOME, one finds that for isoleucine, the three codons have frequencies of 0.47, 0.46, and 0.07. For the six leucine codons, the frequencies are 0.11, 0.11, 0.10, 0.10, 0.03, and 0.55. 

So, Kirk, you’ve done nothing but demonstrate that you don’t understand the concept of codon usage bias.

“Thus: 10/29 = ~34.5% of the total
codons are identical, a number that lies within the 33% - 40% range I described.”

So what? The whole genome tells a different story.

“It looks like codon-usage bias and mutation rates didn’t affect the percentage of identical codons as drastically as you envision.”

Only when you quit looking after n = 1. That’s a cowardly way to go, but it’s entirely predictable.

“Your arguments might look fine on paper, but I’m basing my arguments on biological reality.”

Generalizing from n=1 is a sure way to avoid reality.

“You can argue all you want regarding codon-usage bias and mutation rates, but simple experiments in bioinformatics like the one described above answer your arguments.”

News flash, Kirk: codon-usage tables come from bioinformatics, so your distinction is bogus, as is your conclusion.

Try Googling the phrase “An expanded codon table showing the relative frequence that different codons are used in E. coli genes is shown below.” 

Kirk Durston - #65081

September 27th 2011

Just to clear things up, I am not Genomicus. I find it hard enough to find time to post under my own name, forget about finding even more time to also post under a different name. Genomicus seems to be doing just fine without me and I have not the faintest idea who he is.


Genomicus - #65122

September 27th 2011

John:

>> No, we won’t, because if we tabulate ALL the coding sequences for E.
coli, we will see that the usage of a particular codon (for multiple
codons encoding the same residue) varies between. <<

I am well aware of that, John. I am perfectly aware of codon-usage bias. Do you even understand my argument? The reason the convergent evolution hypothesis doesn’t predict exactly one number is precisely because of codon-usage bias. I’ll be generous and even extend the range from 33% - 50%, instead of 33% - 40%. Therefore, the hypothesis that the sequence similarity shared among FlgBCEFGK is the result of convergent evolution (because of similar functional constraints on each protein) predicts that of all the codons that code for identical residues, about 33% - 50% will be identical—no more, and no less. On the other hand, the Darwinian model predicts no such thing (it doesn’t even make a prediction in this area).

Now, you can argue all you want about how mutation rates et al. affect this prediction, but the fact is that my argument is built on the foundation of biological reality.

Your problem is that you don’t understand my argument. I don’t deny codon-usage bias at all, contrary to what you seem to be implying.

That said, Dr. Durston is correct: yours truly is an entirely different person than Dr. Durston.

<input id=“gwProxy” type=“hidden”><!—Session data—><input id=“jsProxy” type=“hidden”><div id=“refHTML”>


John - #65270

September 29th 2011

“Do you even understand my argument?”

Yes. I understand that it is an argument and not a scientific hypothesis that one would use in seeking the truth.

“The reason the convergent evolution hypothesis doesn’t predict exactly one number is precisely because of codon-usage bias.”

So show how you PRECISELY accounted for codon usage. You won’t.

“I’ll be generous and even extend the range from 33% - 50%, instead of 33% - 40%.”

This isn’t a political negotiation. Simply calculate the range.

“Therefore, the hypothesis that the sequence similarity shared among FlgBCEFGK is the result of convergent evolution (because of similar functional constraints on each protein) predicts that of all the codons that code for identical residues, about 33% - 50% will be identical—no more, and no less.”

No, the hypothesis assumes that natural selection doesn’t operate on silent bases within codons. That is simply false. 

Why have you ignored mutation rates? My questions are simple ones.

“On the other hand, the Darwinian model predicts no such thing (it doesn’t even make a prediction in this area).”

What straw man are you offering as this alleged “Darwinian model”? Aren’t non-Darwinian mechanisms just as important in testing your hypothesis?

“Now, you can argue all you want about how mutation rates et al. affect this prediction,…”

I’m simply pointing out that you’re ignoring them.

”… but the fact is that my argument is built on the foundation of biological reality.”

No, not even close.

“Your problem is that you don’t understand my argument. I don’t deny codon-usage bias at all, contrary to what you seem to be implying.”

If you’re not denying it, simply and precisely show how you have included it in your calculations.

You won’t.

Genomicus - #65333

October 2nd 2011

<!—[if gte mso 9]><xml>

Normal0

</xml><![endif]—>

John:

 

 Well, my arguments actually are supported by biological observation – despite your hand-waving of the evidence. I have indeed taken codon-usage bias into account. My argument is supported by one simple biological observation: namely, that if one aligns two proteins that are not homologous to each other, and if one finds the number of identical codons out of the total codons that encode identical amino acid residues, one will find that the percentage of identical codons is 33% - 50%—no more and no less. On the other hand, this is not true for proteins that share an evolutionary relationship with each other. In that scenario, on average, the percentage of identical codons will fall somewhere outside the range of 33% - 50%. (contd.)

   

/span>

<input id=“gwProxy” type=“hidden”><!—Session data—><input id=“jsProxy” type=“hidden”></p>

<input id=“gwProxy” type=“hidden”><!—Session data—><input id=“jsProxy” type=“hidden”><div id=“refHTML”>
Genomicus - #65334

October 2nd 2011

John:

The above
statements are statements built upon the foundation of biological reality,
whether you like it or not. I could cite a large number of examples to support
my position, and here I will cite several of these examples. Keep in mind that
the following examples are alignments of proteins that are not homologous. In
other words, their sequence similarity is the result of chance, not
common descent.

 

We begin with the already mentioned proteins YscJ and YscT. “They are different proteins that have different ancestries. If one aligns YscJ and YscT (using ClustalW) there will be 29 identical positions. If one threads the amino acid alignment into a codon alignment (using Pal2Nal), one finds that, out of 29 total codons that encode identical residues, only about 10 of those codons are themselves identical.”

 

Thus: 10/29 = 34.5%

 

Our next example is the proteins RecA and dnaC. When these two proteins are aligned (using ClustalW), one finds 38 identical positions, not counting methionine (since, in general, only one codon codes for methionine). Out of 38 codons, 18 are identical.

 

Thus: 18/38 = ~47.4%

 

What about the protein enolase and the protein TolA? Both are completely unrelated to each other and are not homologous. When these two proteins are aligned, 60 positions are identical. Out of 60 codons, a total of 20 codons were identical.

 

Thus: 20/60 = 33.3% 

 

Furthermore, when one aligns the protein agmatinase and the protein SecY, one finds that there are 48 identical residues. And out of 48 codons, 21 are identical.

 

Thus: 21/48 = 43.75%

 

Our final example is the proteins FliZ and FliS, flagellar proteins that do not share homology. When aligned, there are 21 identical residues. Out of 21 codons, 9 of them are identical.

 

Thus: 9/21 = 42.9%

 

    From these five examples, you can clearly see that the percentage of identical codons we would expect from coincidence and codon-usage bias et al. virtually always lies in the 33% - 50% range. I challenge you to provide a single counter instance. If you cannot do that, then you are refusing to engage the evidence.

 

    Now, I mentioned that “this is not true for proteins that share an evolutionary relationship with each other. In that scenario, on average, the percentage of identical codons will fall somewhere outside the range of 33% - 50%.” Numerous examples could be cited here to support this thesis. FliQ and YscS share an evolutionary relationship, yet 26.7% of the codons encoding identical positions are themselves identical. This figure of 26.7% does not fall within the range of 33% - 50%. Moreover, FliK and YscP also share an evolutionary relationship, yet 25.8% of the codons encoding identical positions are themselves identical.

 

   And so it is quite obvious that if we find that the percentage of identical codons encoding identical residues among FlgBCEFGK falls within the range of 33% - 50%, the hypothesis that the sequence similarity shared among those proteins is the result of convergent evolution rather than common descent is strengthened. But if the percentage falls outside of this range, then our suspicion is weakened.

 span style=”“>  In your forthcoming reply, John, please engage the actual evidence I have presented here (and don’t forget to come up with that counter instance I asked for).

<input id=“gwProxy” type=“hidden”><!—Session data—><input id=“jsProxy” type=“hidden”><div id=“refHTML”>
Ashe - #65394

October 4th 2011

Usually only parts of proteins are homologous, not the entire protein (ie domains), and the flgBCEFGK proteins have shared domains. So I don’t think doing comparisons like that is very useful. For example, when I blast FliZ I see some homology with phage integrases but none with FliS. This is not the case with the rod proteins, for example.


Genomicus - #65397

October 4th 2011

Ashe:

You’re being just a touch vague (or maybe it’s me). Would you care to elaborate? Thanks.
<input id=“gwProxy” type=“hidden”><!—Session data—><input id=“jsProxy” type=“hidden”><div id=“refHTML”>
Ashe - #65398

October 4th 2011

I recommend this book on domains, sequence homology, etc:


a href=“http://www.ncbi.nlm.nih.gov/books/NBK20260/” target=“_blank”>http://www.ncbi.nlm.nih.gov/books/NBK20260/

Genomicus - #65399

October 4th 2011

Ashe:

Yea but I’m not sure if you understand my argument. I am arguing that the sequence similarity shared among FlgBCEFGK is the result of convergent evolution, rather than common descent. Sequence analysis reveals that the percentage of identical codons (versus non-identical codons) encoding identical residues between two proteins that share their similarity because of convergence or chance lies around 33% - 50%. This is not true for proteins that share their sequence similarity because of common descent.
<input id=“gwProxy” type=“hidden”><!—Session data—><input id=“jsProxy” type=“hidden”><div id=“refHTML”>
Ashe - #65400

October 4th 2011

Yes I understood that. It just doesn’t make much sense to me. The smallest unit of homology is the domain, not the protein. 


Genomicus - #65401

October 4th 2011

Quite true. But I fail to see the relevance of this? It is my conviction that the homology shared among the similar *domains* of FlgBCEFGK is the result of convergence. <input id=“gwProxy” type=“hidden”><!—Session data—><input id=“jsProxy” type=“hidden”><div id=“refHTML”>
Ashe - #65403

October 4th 2011

Do you realize that homology means divergent evolution from a common ancestor, not convergence?


John - #65445

October 7th 2011

“Well, my arguments actually are supported by biological observation – despite your hand-waving of the evidence.”

Asking you simple, biological questions that you can’t answer is no “hand-waving of the evidence.”

“I have indeed taken codon-usage bias into account.”

No, you have not, as evidenced by your failure to answer the challenge, “So show how you PRECISELY accounted for codon usage. You won’t.”

Your behavior is very predictable.

“My argument is supported by one simple biological observation: namely, that if one aligns two proteins that are not homologous to each other, and if one finds the number of identical codons out of the total codons that encode identical amino acid residues, one will find that the percentage of identical codons is 33% - 50%—no more and no less.”

Aligning proteins that are not similar is an oxymoron.

“On the other hand, this is not true for proteins that share an evolutionary relationship with each other. In that scenario, on average, the percentage of identical codons will fall somewhere outside the range of 33% - 50%. (contd.)”

You have only presented anecdotal evidence, but you are claiming a global conclusion.

“The above statements are statements built upon the foundation of biological reality, whether you like it or not. I could cite a large number of examples to support my position,…”

But see, your insistence on claiming a position as your own is an admission that you are doing politics, not science. It is blinding you to a great deal of biology.

”… and here I will cite several of these examples. Keep in mind that
the following examples are alignments of proteins that are not homologous. In other words, their sequence similarity is the result of chance, not
common descent.”
 
“What about the protein enolase and the protein TolA? Both are completely unrelated to each other and are not homologous. When these two proteins are aligned,…”

You’re not making any sense. 

“From these five examples, you can clearly see…”

…that you:

1) Have no idea how to mathematically consider codon usage bias, and
2) Have no idea what I was talking about when I referred to a non-Darwinian mechanism that you are ignoring. Your ego prevents you from learning more.

”… that the percentage of identical codons we would expect from coincidence and codon-usage bias et al. virtually always lies in the 33% - 50% range.”

FIVE CASES allows you, with an assist from your massive ego, to conclude that something is VIRTUALLY ALWAYS the case?

Hilarious!

“I challenge you to provide a single counter instance. If you cannot do that, then you are refusing to engage the evidence.”

The evidence from the entire genome shows that when an E. coli gene encodes a proline residue, the four codons have the following frequencies:

CCG 0.52
CCA 0.19
CCT 0.16
CCC 0.12

Or take leucine:

TTG 0.13
TTA 0.13
CTG 0.50
CTA 0.04
CTT 0.10
CTC 0.10

These are frequencies for the entire genome, not cherry-picked cases dishonestly presented as examples.
 
“Now, I mentioned that “this is not true for proteins that share an evolutionary relationship with each other. In that scenario, on average, the percentage of identical codons will fall somewhere outside the range of 33% - 50%.””

But you’re too lazy (and scared) to calculate an average.

“Numerous examples could be cited here to support this thesis.”

You haven’t shown that any are real examples or that you have calculated an average.

“FliQ and YscS share an evolutionary relationship, yet 26.7% of the codons encoding identical positions are themselves identical. This figure of 26.7% does not fall within the range of 33% - 50%. Moreover, FliK and YscP also share an evolutionary relationship, yet 25.8% of the codons encoding identical positions are themselves identical.”

So what? You’re ignoring both mechanisms that drive them to be similar AND mechanisms that drive them to be different. These mechanisms can be observed in real time.

beaglelady - #65447

October 8th 2011

John,  you absolutely crack me up.  Peddlers of dishonesty don’t stand a chance when you’re on board.


Genomicus - #65404

October 4th 2011

Make that “sequence similarity.”<input id=“gwProxy” type=“hidden”><!—Session data—><input id=“jsProxy” type=“hidden”><div id=“refHTML”>
Ashe - #65412

October 5th 2011

Right, so the general rule is that proteins that share significant sequence similarity in their domains are homologous because two proteins that perform the same or similar functions don’t necessarily need to share that much similarity in their domains. There are quite a few examples of this in the book I referenced. So invoking convergence with proteins that share significant sequence homology in their domains is at best, ad hoc. 


Genomicus - #65440

October 6th 2011

Ashe:

>> Right, so the general rule is that proteins that share significant
sequence similarity in their domains are homologous because two proteins
that perform the same or similar functions don’t necessarily need to
share that much similarity in their domains. <<

Yea, but the sequence similarity shared among those proteins is mostly found in the N and C termini. This is exactly what would we expect if the observed sequence similarity is the result of functional convergence, since the N and C termini of these proteins are the regions of the portions that interact with the other proteins. In other words, this is where the bulk of the biological function takes place, and so this is where we would expect the functional convergence to occur.


<input id=“gwProxy” type=“hidden”><!—Session data—><input id=“jsProxy” type=“hidden”><div id=“refHTML”>


Ashe - #65444

October 7th 2011

Shared functional domains are common in homologs. What would convince me of functional convergence in this case is if they share very limited sequence similarity and yet perform the same/similar function. 


Page 3 of 3   « 1 2 3