ligand protein interactions for trajectory with united atoms

Thank you for creating this repository!

I am working with a united atom trajectory which causes issues at the the fp.run method:

lig = u.select_atoms("resname LIG")
prot = u.select_atoms("resnum 3-200") 
fp = plf.Fingerprint()
fp.run(u.trajectory[::10], lig, prot)

Which gives the assertion error:

No hydrogen atom could be found in the topology, but the converter requires all hydrogens to be explicit. You can use the parameter ''NoImplicit=False'' when using the converter to allow implicit hydrogens and disable inferring bond orders and charges. You can also use ''force=True'' to ignore this error.

I have confirmed that the plf.Molecule.from_mda method works with the NoImplicit=False and gives the correct topology when viewed in Jupyter notebook:

pmol = plf.Molecule.from_mda(prot, NoImplicit=False)
lmol = plf.Molecule.from_mda(lig, NoImplicit=False)

I hope I didn’t missed something obvious! but how do I use the converted molecules with fp.run?, it seems to expect an MDAanalysis atom group does not seem to have kwargs during plf.Fingerprint() or fp.run

Thanks in advance,

Ali

[askusay]

Hi Cedric,

Happy to report that I’ve now solved this issue using PyMOL, I did it by adding missing hydrogen atoms to all simulation frames.

How to:

1. Load topology i.e. pdb file into PyMOL
2. Select the ligands (lig sel) and set the valence of their bonds using the valence guess, lig_sel command. (see notes for some tips on this)
3. Load the simulation using load_traj
4. Run the h_add (see notes)
5. Save the simulation preferably as a dcd file using save_trj or as individual pdb files - much slower and larger files. The save_trj function must be installed from here
6. Load simulation using MDAnalysis and run uni.atoms.guess_bonds() where uni is the universe object. More on this in #25
7. Now the analysis should work as intended

Notes:

1. I suspect that the valence command may not work for complex ligands, in this case a combination of the valence and bond with the appropriate options may be needed
2. Some elements types may need to be manually set in PyMOL, I.e. to set an atom to bromide: alter atom_sel, element="BR". This can also be done in MDAnalysis by iterating over the atoms and setting their elements
3. The valence must be set correctly before running h_add otherwise hydrogens will be added incorrectly i.e. for aromatic systems

Thanks for your help Cedric! Feel free to close the ticket or provide any comments on this.

[cbouy]

Hi Ali,

Good to know that you’ve managed to solve the issue! Thanks for posting the procedure that you’ve used, I’m sure that it will be helpful to someone.

Take care, Cédric