Separate Transformations for SepTop

[mathilfiker]

Thank you for all your work and dedication to this project,
Is your feature request related to a problem? Please describe.
I really like the fact that the HybridTopologyProtocol allows you to save the complex leg of the cycle and the solvent one separately. This makes it easier to schedule the partitions for submission to HPC environments, and to experiment new approaches in one environment only.

Describe the solution you'd like
Is it possible to do the same with the SepTop protocol? I read in the user guide that "both legs of the thermodynamic cycle are constructed and run sequentially in the same ProtocolDAG", is it possible to force this?

Additional context
Extra question: in your implementation, is the ouput of the solvated leg just the relative hydration free energy? In the publication referred in the user guide, it is mentioned that the two quantities coincide, but I see it leass clearly in the depicted thermodynamic cycle.

[IAlibay]

Is it possible to do the same with the SepTop protocol? I

Unfortunately not at this time (the reason turns out to be the answer to your second question). In the near future, we will hopefully be moving to a new execution engine (exorcist) that will allow you to parallelise the different execution units, which would address some of your use case.

If you wish to experiment with just one leg, the best approach would be to use the python API directly. There you can "manually" execute different portions of the transformation. We can provide more details on this, if it's of interest.

Extra question: in your implementation, is the ouput of the solvated leg just the relative hydration free energy?

Not as far as I know (cc @hannahbaumann - please correct me). We use a partial charge annihilation scheme for our transformations, so that leg of the transformation is missing the gas phase recharging step.

[mathilfiker]

I would like to experiment on one leg only, could you please provide more details on this?

Further, if I do not define a protein in the system prepration but just ligand and solvent, can I use the same protocol to run relative hydration free energies or would that break?

[IAlibay]

I would like to experiment on one leg only, could you please provide more details on this?

We're out of office at the free energy workshop right now, but we will post the details here soon.

Further, if I do not define a protein in the system prepration but just ligand and solvent, can I use the same protocol to run relative hydration free energies or would that break?

Unfortunately not, SepTop doesn't do the gas phase sampling necessary for RHFEs.