[QUESTION] Area per lipid in membrane-protein complexes

[pbuslaev]

What is the topic of your question: Usage

Add your question below: Is there a pipeline to compute Area per lipid for lipids in membrane-protein complexes?

[pbuslaev]

My workaround is to include CA atoms in lipid_sel, but I think it is not the perfect way. I believe that efforts around #63 can be quite useful and I can happily help with that if needed.

[p-j-smith]

Hi @pbuslaev, I'm afraid there's no pipeline for this at the minute.

My workaround is to include CA atoms in lipid_sel, but I think it is not the perfect way.

Yeah that would get you a value, but you're right it's not ideal

I believe that efforts around #63 can be quite useful and I can happily help with that if needed.

That's a good idea. Are you thinking you'd find the protein atoms neighbouring a lipid, then include those in the selection for the area per lipid?

I can happily help with that if needed.

Any help would be much appreciated!

[p-j-smith]

I just realised that the lipid-protein neighbours might change at each frame, and there's no way to change the lipid_sel for each frame in the area per lipid analysis (the selection can be an updating one, but there's no way to update the selection itself).

Thinking again, you might be able to use AssignCurvedLeaflets and set midplane_sel='name CA' and an appropriate midplane_cutoff value. I think this would essentially do what you proposed - find CA atoms near to lipids and include these in the leaflets.

[pbuslaev]

That's a good idea. Are you thinking you'd find the protein atoms neighbouring a lipid, then include those in the selection for the area per lipid?

My thought is this selection can be automatically selected to find area that is occupied by membrane for every frame. Thus, if you add not_membrane_selection field to area_per_lipid function, the algorithm can use this selection to find atoms in lipids+neighbors set that belong to not_membrane_selection and exclude those from area calculation