search

choderalab / perses

Experiments with expanded ensembles to explore chemical space
http://perses.readthedocs.io
MIT License
179 stars 50 forks source link

Integrate new atom mappers #1204

Open ijpulidos opened 1 year ago

ijpulidos commented 1 year ago

We should be able to integrate other atom mappers such as the TimeMachine mapper or similar, that we could use in perses for doing the mapping.

ijpulidos commented 1 year ago

This requires refactoring the objects so that they accept atom maps instead of generating them on the fly.

ijpulidos commented 1 year ago

Wwe could also try https://github.com/michellab/BioSimSpace/blob/devel/python/BioSimSpace/Align/_align.py#L719

caveat: I think it uses interconversion of the RDKit mol to a SireMol to get to the atom mapping. This may not be ideal

IAlibay commented 1 year ago

@ijpulidos we're doing a lot of work on the openfe side to add new mappers (lomap, kartograf, and others eventually). Would it not make more sense for perses to instead directly plug into those mappers?

ijpulidos commented 1 year ago

@ijpulidos we're doing a lot of work on the openfe side to add new mappers (lomap, kartograf, and others eventually). Would it not make more sense for perses to instead directly plug into those mappers?

Yes! That makes a lot of sense.

ijpulidos commented 1 year ago

Time machine mapper is in https://github.com/proteneer/timemachine/blob/master/timemachine/fe/atom_mapping.py

IAlibay commented 1 year ago

cc'ing @richardjgowers

richardjgowers commented 1 year ago

@ijpulidos you should probably think about Licenses if you're looking in BSS. And yes if you refactor so that the AtomMapping is an input then it will be relative easy to get this from Lomap/Kartograf/etc, I think wrapping up TimeMachine's approach is interesting, it's a more modern look at what rdFMCS can do compared to other options.