Allow for multiple alchemical waters

[IAlibay]

From today's Perses-OpenFE call - it might be good, once we've finished up validating the alchemical water code, to see if we can support net charge changes > 1.

[zhang-ivy]

In theory, mutations from positive to negative charge (or vice versa) would be good examples for this, e.g. GLU to ARG. There aren't any such mutations in the protein:protein benchmark set (i.e., barnase:barstar) I've been working with, so we wouldn't have any experimental data to compare to, but could do some cycle closure/convergence tests on these types of mutations in simpler systems first (like terminally-blocked amino acids) and then more complex systems like barnase:barstar.

I don't have any small molecule transformation examples off the top of my head, but there may be some mutations with charge change > 1 in kinase:inhibitor datasets (tagging @sukritsingh for input here)

[sukritsingh]

There are a couple in Abl kinase off the top of my head - I'll link to potential papers below but mutations in Abl kinase that come to mind off the top of my head are:

1. E255K
2. E275K
3. E279K
4. E281K
5. K419E
6. E450K

Thoughts/considerations here:

1. It may be useful to pick a couple of examples in either direction as test cases (ie one going from negative-> positive and another going from positive-> negative).
2. All the major charge change mutations involve GLU and LYS - is it worth trying other residues out as well?

Pasting relevant papers here that are mentioned in FEFlow issue #7

https://www.nature.com/articles/s42003-018-0075-x https://pubs.acs.org/doi/full/10.1021/acscentsci.9b00590 https://pubs.acs.org/doi/full/10.1021/acscentsci.8b00717#

[IAlibay]

Thanks @zhang-ivy and @sukritsingh

I need to review the above papers (so apologies if the answer is already there). One immediate thought here is whether or not a > 1 net charge transformation is actually a reasonable single step perturbation. Are there any cases here where going through a neutral residue as a middle point is known to either not be feasible or bring along some significant disadvantages?

[sukritsingh]

Oop sorry for not seeing this till just now!

Are there any cases here where going through a neutral residue as a middle point is known to either not be feasible

Sorry just to clarify - are you asking whether or not there are examples of mutations with >1 net charge transformation? If so, then the answer is yes, there are plenty of real-world examples of those. If we were going through a "neutral" residue then the real world experiment for that would be going through 2 mutations (Addendum: While possible to occur, this kind of pathway is less likely to occur in biology and if there was a selection pressure to achieve a >1 net charge mutation then that would occur directly. If a neutral residue appears mid-pathway in biology then, for our purposes, it had some degree of selection/value by itself.)