User-defined dihedral angles for clustering?

[vvoelz]

The Voelz lab is currently building MSMs from simulations of non-protein small molecules and peptidomimetics that don't have the usual phi/psi/chi/omega, etc.

It would seem easy enough to add user-defined dihedral definitions by passing in a flat file of listing quartets of atoms (specified by residue number and atom name) -- this is something we have gotten to work before with an earlier hack. Maybe this would be a useful feature in general?

[rmcgibbo]

Internally, there's an n_dihedrals x 4 collection of the atom_indices of the atoms which form each of the dihedrals. The easiest thing would be to let that get passed in as a flat text file, analogous to the AtomIndices.dat file for RMSD. Would that be a little more general-purpose than a residue/atom name file?

[kyleabeauchamp]

Robert: I wonder if we should make the library-style usage of MSMBuilder easier. Right now, our Clustering scripts do everything, but it's pretty hard to replicate those functions in library calls.

For example, I imagine small molecules and peptidomeimetics will leave to a lot of edge cases to consider--it could be better to make the library easier rather than the scripts more general.

[rmcgibbo]

@kyleabeauchamp: That's easy to say -- the hard part is specifying what you mean by "easier". To some extent it's just complex because there are a lot of options, a lot of functionality. That's not to say it couldn't be better.

Whereas, we both know exactly how we could add the dihedral_indices flag, what the code would look like, and where it would fit in, so that's actually easy.

[rmcgibbo]

@kyleabeauchamp: Another thing would just be to have a full collection of well commented IPython notebooks that replicate / extend most of the script functionality, but make it kind of "inviting" to modify them.

[vvoelz]

@rmcgibbo: I agree -- passing in a DihedralIndices.dat file would be easiest. It could go like this: Cluster.py dihedral -a user -f DihedralIndices.dat hybrid .... Or maybe you have better ideas?

@kyleabeauchamp: Several people in my lab (myself included) have modified the library as you suggest. But this is something that we do all the time so it seems like it should be a standard feature

[kyleabeauchamp]

I agree this sounds like a useful feature.

[vvoelz]

How about I make the modifications and get them working in a branch, and submit a pull request? I'm still getting the hang of git, so it will be some good practice for me.

[kyleabeauchamp]

Great! On May 23, 2013 4:50 AM, "Vincent Voelz" notifications@github.com wrote:

How about I make the modifications and get them working in a branch, and submit a pull request? I'm still getting the hang of git, so it will be some good practice for me.

— Reply to this email directly or view it on GitHubhttps://github.com/SimTk/msmbuilder/issues/196#issuecomment-18338364 .