I would be interested in the data and equations he is using to get those numbers.
I would also be interested in why @gpuccio expects transitional sequences between living species. If he were a linguist he would be claiming that French, Italian, Spanish, and Portuguese must have popped into being as separate languages because we can not find intermediate modern languages that link them together to a common Latin tongue.
Thereâs nothing respectful or ethical about your behavior in this matter. If you misrepresent someone, intentionally or not, you correct the record. Only then do you move on.
Itâs amazing that so many who deny evolution misrepresent it in this way. They do this for both genes/proteins and species.
Yep.
Curtis, while some of the concepts that underlie the approaches taken to estimate FI may have some use, I personally donât have much use for FI. It doesnât help me in studying protein function or even evolution, and it isnât much use for teasing out RNA structure, function, or evolution.
Off the top of my head, what I may use some of the concepts for is estimating whether or not a particular selection or genetics screen may be fruitful. It does help to relate expectations with things like library or population size. But I am not going to do the âFIâ thing to figure these matters out.
I beg to differ. @gpuccioâs mainstay - subtracting the bit scores from BLAST-based pairwise alignments - is nothing like the approach taken by Hazen et al.
He has a method. The blast score is only a piece of that method. The paper has a definition. His method uses the definition in the paper. I am not sure the point you are making but non the less lets wait for him to finish his projects and return for round 2.
I raised this issue much earlier in the discussion at PS of gpuccioâs arguments. Gpuccio says that he is using the definition as given by Hazen, Carothers, Griffin, and Szostak (2007). Their discussion assigns to each sequence a degree of function Ex. However gpuccio actually does not do that but further restricts himself to talking about cases where many sequences have no function (Ex = 0) and some have function (Ex > 0). If the fraction that have no function is big enough, that creates a problem for having natural selection find its way to the functional sequences.
Except, in a long thread at TSZ, gpuccio ended up by
acknowledging that he also meant to include cases where many sequences did have Ex > 0. But he felt that the amount of function they had was too little for natural selection to be affected by it.
So which definition it he using? HCGS? His own where Ex = 0 for most sequences?
That is categorically false. Rather @gpuccio claims his method approximates FI as defined by HS, and he makes this claim with out any valid theoretical justification. His method is not guided by his definition of FI in any way and, in fact, does not estimate it.
The theoretical basis is that preservation indicates functional constraint. He is guided by the FI definition in H and S. Unless someone is willing to understand one of his ops and make specific criticisms then this conversation has no value. I cited one of his articles to John above.
There are many cases where the concept of natural selection as Joe is commenting above is not likely present as in prp8 where the original version starts at greater than 500 human functional bits and where accuracy is required to maintain life.
HI T and @Rumraket the comments to read are at 651, 654 and 663. My concept and his agree I make a clean separation by separating functional bits from functional information.
Once again you confuse your unsupported laymanâs opinion with fact. You do that with almost every post.
What we get here is high level concepts. If you claim selection is a method based on a historic observation then show based on the data where it happened. Guys like Joe Thornton have done this with enzymes. You need function before you have a chance at selection. When we see chromosomes in eukaryotic cells where is the evidence (other than speculation) that selection was involved. Where is the evidence it can direct a sequence toward complex function versus away from it.
Minds can select directly against a specific sequence and not a generic selection based on the environment.
LOL! So now you claim science has no evidence natural selection happens? 
Then you still wonder why people question your knowledge of basic evolutionary biology.
You still havenât shown us how minds alone can physically manipulate matter. Itâs sure bet you never will.
Catalytic antibodies, for one.
There are many others that have been presented to you. Why do you keep repeating questions that have already been answered?
@colewd, this is exactly what I did. Can you enumerate the technical critiques I explained to him?
If you canât, please stop spreading confusion. If you can, tell me precisely on what basis you dispute them. If you canât even enumerate what my specific criticism are, you need to slow down, retract that claim, and catch up.
Thats not the claim.
Fair enough. I will re read the post and seek clarification of your position. No one is claiming his measurements are perfect. The question for me is are they a starting point for further research as we can never measure directly the entire sequence space of a long complex protein.
We discussed how it could be improved and extended, but that his analysis was invalid at this time. @gpuccio was not interested in improving or extending his analysis, so he does not share the your desire to further his research.
I am glad to hear you acknowledge that we cannot directly measure the entire sequence space of a long complex protein. That is exactly correct, and why we canât make confident claims about design based on sequence space.
We even produced an example of a paper that measured FI increases over time. He did not appear to care much about reading it.
Preservation does indicate functional constraint. It does not, however, provide an estimator of FI as defined by @gpuccio. What it provides is an estimate (not necessarily a very good one) of the area of the fitness peak that the sequence occupies â i.e. the number of sequences that are equally functional and that can be reached by (more or less) neutral mutations. It does not provide an estimate of the total target space of functional sequences, either in the neighborhood of this peak but with lower fitness, or in the search space more generally.
In addition, in the safe analogy, he entirely drops the definition of FI in terms of target and search space, and instead starts treating it as measuring how long a path is required in search space to reach a target, which is quite a different beast.
If we could actually estimate FI as defined â if we knew how many targets there were for a given function â it would at least tell us something about evolution, specifically the probability that the function in question would be found by randomly assembling proteins. That would be interesting, even though it still wouldnât tell us whether the function resulted from evolution.