Separated Topologies method

[nffaruk]

Hi, nice work on the package! As you note in your paper, computational resource limits can get in the way of practicality. I was wondering if you considered the Seperated Topologies method here: https://pubs.acs.org/doi/10.1021/acs.jctc.3c00282 How difficult would it be to support such a scheme within a3fe and if you have pointers on which source files to look at to modify. Their repo: https://github.com/MobleyLab/SeparatedTopologies

[fjclark]

Hi, thanks! And thanks for the question/ suggestion. It's an exciting approach and I'd like to play around with it. The best way to implement this in a3fe would likely be:

1. Modify BioSimSpace to allow setup of SepTop calculations in GROMACS (as a3fe relies on BioSimSpace for all setup)
2. Update a3fe to handle GROMACS output files

I don't think 2. should be too hard, and I'm hoping to get round to it in a few months to allow ABFE with GROMACS. Unfortunately, 1. might require more time than I'll have before I finish my PhD (early next year), although I'll have a closer look at the Mobley Lab repo you sent.

[nffaruk]

Thanks for the response. I've found out that OpenFE is in the process of incorporating SepTop, so then it may not be worth it considering your time crunch till graduation.

[fjclark]

Great, that's exciting. Yes, in that case I won't put time into this soon, but I'll keep the issue open as an enhancement request for future. Thanks.

[lohedges]

There is also femto, which supports both AToM and SepTop via OpenMM. This can be used as a standalone library, or via the command-line. (It also has support for running using repex and analysis.)

[nffaruk]

Thanks for pointing to femto, will check it out! But from a glance, a3fe seems a lot more streamlined than it - will be cool if the enhancement ever comes to fruition.