How to use pyautofep calculation when the ligand is peptide

[CPU-HORDB]

[luancarvalhomartins]

Dear @CPU-HORDB,

Thank you for you message. I am moving it to the discussions, as I think it is more suitable to that.

[luancarvalhomartins]

Originally, PyAutoFEP was not written for peptides, although I agree it would be quite useful.

For using it as is, the complexity will depend on what parameters you want to use.

• The fast and dirty way would parameterizing the peptide as a small molecule the using the regular workflow. It should work, but the parameters will probably be worse than the FF parameters.
• The correct way would be to obtain parameters for the peptide using the FF you will use and than pass that topology as ligand topology. You may experiment with gmx pdb2gmx and gmx grompp -pp, then trimming the resulting topology to keep only the peptide parameters.

Note that obtaining an MCS will be challenging. Maybe you want to pass mcs_custom_mcs. In case you are inverting stereocenters, it will be extra challenging and I would guess find_mcs_3d will fail. In case there are no inversions, maybe find_mcs will work. Superposition may work, it is hard to know before hand. The larger the peptide, the harder the superposition. In case superposition fails, LMK. I could try coding the option to use input poses as pose.

Feel free to keep in touch regarding this.

[ninomiya829]

Dear @luancarvalhomartins,

I've recently embarked on using PyAutoFEP to estimate the FEP for a cyclic peptide comprising roughly eight amino acids, along with its derivatives with related proteins.

However, after executing the command, generate_perturbation_map.py --map_type=star --map_bias=1918_lig --input lig_data/1918_lig.mol lig_data/1919_lig.mol lig_data/1920_lig.mol

I've noticed that the program seems to be stuck at a certain point for an unusually long time.

It seems stucking in mcs_data = os_util.starmap_unpack(find_mcs, thread_pool,[[[molecules[mol_i], molecules[mol_j]], None, verbosity]for (mol_i, mol_j) in todo_pairs],itertools.repeat({'completeRingsOnly': True, 'matchValences': True,'ringMatchesRingOnly': True})) step

It's got me wondering if PyAutoFEP might not be the ideal tool for handling such a large cyclic peptide, or perhaps I've overlooked something in my setup.

Would you happen to know if there's a workaround or a particular adjustment I could make to tackle this issue? I'd really appreciate your insights and assistance on this. Thanks a bunch for taking the time to help me out!