The link to "Protein interactions aren't easily generated" is a link to you stating that "I don't think that it is necessarily true that protein-protein interactions are generated simply," followed by your linking to a paper that has nothing whatsoever to do with protein-protein interactions.

No. What I said was "protein interactions aren't easily generated", I then linked to a post that links to a paper that shows how a single change is not sufficient to generate Na⁺ binding and activation (a single change is sufficient to abrogate it). This suggests that protein-protein interactions aren't necessarily easily generated either.

genomic:

2) If it is difficult to generate protein-protein interactions, wouldn't you predict that most proteins should be very soluble in physiological ionic conditions?

I don't see how that follows.

Are you claiming that ubiquitination is specific for lysine residues, or are you predicting that as a hypothesis that you are interested in testing?

We already know ubiquitin forms covalent bonds with other proteins at ubiquitin's C-terminal end; the location of a glycine residue. Lysine side chains of other proteins are part of the bonds. This does not need testing. What assumptions are you making about ubiquitin that is relevant to whatever it is you believe?

So you're offering ubiquitin as an example of a specific enzyme? Are you claiming that ubiquitination is specific for lysine residues, or are you predicting that as a hypothesis that you are interested in testing?

Do the proteins you consider "most basic of proteins" exhibit specific or merely selective interactions with other proteins and substrates?

They are specific, interacting with just one or only a few substrates and they catalyze a specific type of reaction.

I would guess that the selective value of proteins is largely determined by specific protein properties enabling interactions between them.

Your answer has nothing to do with my question, so perhaps my question was unclear–a lot of practicing biologists confuse specificity with selectivity.

When I used the term "selective" in my question, I was not referring to selection in evolution–I was using it as a biochemical term. Selectivity (for protein-protein interactions/enzymes) is well-defined as binding/using a range of partners/substrates, some better than others, and can be measured. Since Guts doesn't seem to be interested in this question, would you mind having a go at making a prediction, assuming I've clarified the question sufficiently?

You also wrote,

I would imagine the improvement would be dramatic but would hasten to point out the limited scope of this example. What if the analysis instead focused on a function involving a score of proteins and improvement was linked to changes in multiple proteins?

Obviously, the analysis would take longer to complete. But wouldn't the amount of time required to develop protein-protein binding in this way be highly relevant to deriving a new function from any existing system that meets the ever-shifting definition of IC?

As a general rule the more basic the protein function, the more difficult it is to detail a plausible pathway to its origin IMO. If this is so then it is the most basic of proteins in IC systems that should be the object of inferences for ID.

But isn't that related to my question (hopefully now clarified) about specific vs. selective interactions? Do the proteins you consider "most basic of proteins" exhibit specific or merely selective interactions with other proteins and substrates?

3) Since you'd probably try to argue that interactions between random proteins aren't specific, are virtually all protein-protein interactions in biology specific or merely selective? Which would you predict from a hypothesis of intelligent design?

I would guess that the selective value of proteins is largely determined by specific protein properties enabling interactions between them.

4) If I replaced a bacteriophage protein, required for binding to a host protein allowing infection, with a random sequence, how many orders of magnitude improvement of phage infectivity would you predict that mutation and selection would produce in a few rounds of infection?

I would imagine the improvement would be dramatic but would hasten to point out the limited scope of this example. What if the analysis instead focused on a function involving a score of proteins and improvement was linked to changes in multiple proteins?

This brings up an issue that has a bearing on irreducible complexity. To illustrate my point I'll use some homemade phrases. IC systems can be thought of as layered. Some are more basic to the overall function of a cell than others. These functions in general could be thought of as having been conserved at the outset. For example, a protein found in a cellular membrane could have a vital function. But its synthesis is dependent on other proteins that are more basic to overall cellular function in that they also are involved in the synthesis of other proteins unrelated to the cellular membrane protein. As a general rule the more basic the protein function, the more difficult it is to detail a plausible pathway to its origin IMO. If this is so then it is the most basic of proteins in IC systems that should be the object of inferences for ID.

Protein interactions aren't easily generated , it's the shape of the protein, the position of active domains, etc that determines its functional interface with other proteins.

That raises several interesting questions:

1) The link to "Protein interactions aren't easily generated" is a link to you stating that "I don't think that it is necessarily true that protein-protein interactions are generated simply," followed by your linking to a paper that has nothing whatsoever to do with protein-protein interactions. What were you trying to convey in both cases? Neither link makes any sense to me. How did you get from thinking that they aren't easily generated to knowing that they aren't easily generated?

2) If it is difficult to generate protein-protein interactions, wouldn't you predict that most proteins should be very soluble in physiological ionic conditions? Is that the case, even for random sequences?

3) Since you'd probably try to argue that interactions between random proteins aren't specific, are virtually all protein-protein interactions in biology specific or merely selective? Which would you predict from a hypothesis of intelligent design?

4) If I replaced a bacteriophage protein, required for binding to a host protein allowing infection, with a random sequence, how many orders of magnitude improvement of phage infectivity would you predict that mutation and selection would produce in a few rounds of infection?

Sorry, I lost track of the thread and forgot the conversation had moved on.