Hi, I want to dock a protein to a TM protein embedded in a lipid bilayer.
Is there a way to do that with an implicit membrane model?
My current approach is to place explicit POPC lipids around the TM protein.
What would be the simplest way to make Rosetta being able to read the pdb file containing POPC?
If you have at least one protein with a transmembrane helix, you should be able to use Rosetta's implicit membrane model to dock them, even if one of the proteins does not embed into the membrane. (see https://doi.org/10.1371/journal.pcbi.1004398 for the current best-practices in working with membrane proteins) You should be able to combine the standard (soluble) protein-protein docking protocols with the setup for membrane protein modeling. (While there's a special application for membrane protein-protein docking, that's for the special case of docking two membrane-embeded proteins, which is a harder case due to the need of keeping both properly oriented in the membrane. If one of the partners is soluble, the standard (soluble) protein-protein docking proceedure can be used, so long as you use an energy functions which adequately accounts for the presence of the membrane.)
These approaches typically work with an implicit membrane model. Typically in Rosetta sampling we don't add any explicit membrane molecules. That's not to say you couldn't do the modeling in the presence of explicit lipid molecules. The trick would be that Rosetta doesn't "read" POPC by default. You would likely need to create parameter files for the POPC molecules using a similar sort of approach (molfile_to_params.py) that is used to setup ligands for ligand docking, though you'll also need the centroid parameters for protein-protein docking. Once you specifiy those parameters on the command line with -extra_res_fa and -extra_res_cen, you should be able to read in PDB with POPC molecules.
However, that's not necessarily a guarantee that they'll be scored appropriately by the (implicit) membrane model. If you can get away with working with an implicit membrane, I'd recommend doing that.
Would that actually mean that one only has to set this?:
MP specific options should be only available to the mp_dock application, or?
How is about the span file? Is that needed or is it just about the z-coordinates?
Just in case somebody else tries something similar.
After some conversation with rmoretti it turned out that it is currently not possible to dock a soluble to a TM protein. (Also not using rosetta_scripts.)
This should be fixed in the near future.