Atomistic membrane?

[jarvist]

For some work we are currently doing, it would be extremely useful to generate both a coarse-grain and atomistic membrane. I suppose 'backward' (https://cgmartini.nl/docs/tutorials/Martini3/Backward/) could be used, but this would still require an atomistic topology.

Is there any more sensible approach?

[MikkelDA]

Hello

I haven't personally used Backward before but i have previously used CG2AT (https://github.com/owenvickery/cg2at) to backmap my Martini3 systems. It might require you to create some mapping files for your lipids but from what i can remember that shouldn't be too difficult. There is a thorough explanation of the program, file layout and how to write fragment mapping files on their GitHub.

Generally i would advice people to use a different program than COBY for atomistic system creation if the system contains membranes, as COBY does not insert different lipid conformations (as CharmmGUI does). COBY just places a number of copies of the same lipid structure with a random rotation around the z-axis (with each bead having had a small random kick applied to it). I do think it would be fine to create a system in M3 with COBY and then backmap it with CG2AT as CG2AT performs its own minimizations and equilibrations, thus "correcting" some of the simplifications done by COBY.

Cheers, Mikkel

[jarvist]

Thank you Mikkel! I'll give CG2AT a spin.
Certainly I think it's fine to use an external tool, but I imagine this is quite a common use case for CODY, so it would be good to document / smooth off the edges so it's obvious how to do it.