PyAutoFEP is it possible to calculate the membrane system at present?

[Teng-Dan]

Hi, it is mentioned in the manual ‘It automates the building of a simple receptor-ligand system and will fail for complex cases (eg: membrane systems, systems with vacuum regions, solvent mixtures, and systems with nucleic acids on cofactors).’ So I want to know whether PyAutoFEP can be applied to the calculation of membrane system. Looking forward to your reply, thanks!

[luancarvalhomartins]

Yes, PyAutoFEP is able to handle membrane systems. The automated system builder, however, cannot. Therefore, you will have to prepare the system and supply it as a pre-solvated structure, ideally, alongside the topology. You will also have to use custom mdp files matching your membrane (semi-isotropic pressure coupling? anisotropic? compressibility?). I am planning to add a tutorial for membrane proteins anytime soon. In case you need further help, don't hesitate to contact me here or via email.

[Teng-Dan]

That's great! In fact, I don't know how to manually process and commit these files when using PyAutoFEP. The protein currently studied in my project is G-protein coupled receptor (GPCR), a very representative membrane protein. Can you consider GPCR as an example when preparing the membrane protein tutorial? Could you please let me know when you update the membrane protein tutorial, I am looking forward to your tutorial, thank you very much!

[luancarvalhomartins]

My plan is to use B2AR in the tutorial. I am a bit swamped right now, but I will do my best to work in this tutorial soon. In case you are interested, I can send you an alpha version of the aforementioned tutorial before commiting a polished version to the repo (feel free to reach me by email if so).

In case you are interested in moving forward regardless of the tutorial, I suggest you to (hopefully, it may also be useful to other users):

• Process the PDB of the membrane protein (missing residues/atoms, cosolvents/crystallization species, protonation)
• Use the OPM server (https://opm.phar.umich.edu/) to get the orientation of the protein
• Prepare a bilipid+water+ions box (eg, a truncated hexagonal box)
• Use gmx mdrun -membed to insert the protein in the membrane
• Minimize, then equilibrate the system for, say, 50-100 ns
• Use the last frame from this equilibration, along with its topology, as input to prepare_dual_topology.py using pre_solvated. Notes:
  • Currently, prepare_dual_topology.py won't run gmx genion for pre_solvated systems, so be aware of charge differences
  • Make sure to include any additional topology files as extrafiles or extradirs
  • In case you equilibrate the system with a ligand, currently, prepare_dual_topology.py requires that the residue name of this ligand to be LIG. I hope to solve this, but it's not high priority.

[Teng-Dan]

That's sound great! Looking forward to your tutorial. I'll try to run as you suggested first, thanks for your reply.