How would you calculate the FI of sequence that didn’t have ubiquitin function? From what I understand, only sequences with function are included in the calculation. If what you describe is true, then @gpuccio and others are going to have include the entire evolutionary history of the ancestral sequences that preceded ubiquitin, or any other functional protein, in their calculations. Needless to say, this means we can’t calculate FI with our limited knowledge of those sequences.
It’s starting to look more like a mine field problem. Sequences go through a random walk in a massive mine field, and when they finally hit a mine we stand amazed at the number of supposedly random events that had to happen just right in order to step on that specific mine. However, at no point in the pre-mine phase would anyone have thought any of those random events were peculiar.
If the final step required 4.3 bits, then so to did every step before it. Each and every step is 4.3 bits. Not 500 bits. 4.3 bits. None of these steps would be considered out of the ordinary for evolution. These steps are only given importance once a function is found, and we all know what this is: The Sharpshooter fallacy.
You are making assumptions that we don’t have data for. What is the real origin of ubiquitin? We don’t know at this point. The methodology is straight forward which is comparing ancestral sequences to human sequences and to see how they change. If the change is minor over long evolutionary periods then were observing functional constraint.
Unless I’m missing a crucial step in Gpuccio’s analysis (in which case, somebody please correct me), this is wrong. No ancestral sequences are estimated or used in the analysis. Only sequences from extant organisms are used.
No. When I point out what I see as an inconsistency in Gpuccio’s numerous different pronouncements and claims, that is not “making up a false narrative”. If I am wrong when I argue that there are problems with Gpuccio’s methodology, then I am just wrong. There is no “false narrative”.
If a single mutation finds a complex function this is impossible unless the rest of the configuration had a workable sequence for the new protein.
Yeah that’s the whole point. That’s why one protein with function A, can evolve into another protein with function B if it is nearby in sequence space. The two functions might technically be separated by as little as one substitution.
The workable sequence is the FI the old protein gives to the new protein.
No, the old protein has no FI for the new protein if the old protein can not do the function of the new protein. None. Zero. The old protein’s sequence is part of the nonfunctional area of sequence space with respect to the new function.
If there was no FI in the prior sequence like a sequence of the same amino acids the scenario you described would not work.
Yes it would work. It would work by definition. That is the method by Hazen & Szostak.
Then the FI of ubiquitin can’t be calculated because you have to include all of the ancestral sequences that preceded ubiquitin, clear back to the first life. This is true of every single protein and gene sequence.
The problem with this idea is that functional constraint is contingent.
The best analogy I have found for this concept is the game of Jenga. If you play a few games you will notice that some of the blocks are loose at the beginning and can be removed with no problem. However, as blocks are put on top those same blocks may become integral to the stability of the tower. That’s how biology works. A new feature may superfluous or just a small asset, but as other proteins evolve to interact with and improve that first protein it starts to become integral. That results in higher conservation of sequence since it has more interactions that it has to keep.
This also shows why the Sharpshooter fallacy is a problem. When an entire pathway was just one gene it could have gone in many directions. Once other proteins evolve and add to that pathway it gets locked in. If you calculate the odds of getting just that pathway you will get a false probability because you are ignoring all of the other pathways that could have evolved. Apply this to ubiquitin and you will start to understand where we are coming from.
Wait, now I’m even more confused. Is Gpuccio reading this thread, replying at UD, and then you’re copying his responses (apparently without indicating that you’re doing so) back in this thread? That seems like an awfully silly way to have a conversation.
Read their paper Bill. A sequence that does not perform the function is by definition outside the threshold of minimum function, and so is just part of the nonfunctional sequence space. It doesn’t matter that it is only one mutation away from being able to perform the function. If it does not perform the function, it is non functional. No matter if it is 99.99999999999999999999% similar to a functional sequence.
By pure definition I agree with you but you are in reality only adding 4.3 bits by adding an amino acid. The balance of the functional information was already in the AA sequence.
If this is not true then you have no way to differentiate a garbage sequence then one that is only 1 AA away from function.
If adding 1 AA or 4.3 bits could change any sequence into functioning ubiquitin then you would have a better case that the mutation was adding more than 4.3 bits of FI.
How would you calculate functional information in this case?
So how would you differentiate between them? What methods would you use? What calculations would you use? If I gave you a random sequence, would you be able to tell me if it was 1 AA away from being functional, and which function it would have with just one change?