According to MEMBPLUGIN defaults (used in our paper (doi:10.1093/bioinformatics/btu037)) and by analogy, selections for phospholipids and sphingolipids consist on the following 3 atoms:
C2 (the middle one) of the glicerol part in phospholipids or the C2 (bound to the amide bond of the acyl-chain 2) of the sphingosine part.
C21: the first carbon of the acyl-chain 2 (the one forming part of the carbonyl group).
C31: the first carbon of the acyl-chain 3 for phospholipids (not sphingolipids).
C3: of the sphingosine part for sphingolipids.
For sterols such as cholesterol, the selection consists on a polar atom of the polar head (O of the hydroxy group).
Keep in mind that modular topologies like AMBER Lipid14 force field, where a single lipid molecule has several residue names, are not curently supported at least for Area per Lipid computations. You will need to rename your molecules if this is your case.
I have no experience with cardiolipins, but probably taking the atom names for PMCL by doing an analogy among two phospholipids and PMCL could be a good approach. This should result into a set of 6 atom names ("C12 CA1 CB1 C32 CC1 CD2" in CHARMM) for the resname "PMCL" .
Maybe, you can add extra atoms that are usually in the same plane and in the rigid region of the lipid. The objective here is that the selected atoms define a membrane plane and the bulkiness of the lipid in that plane. Just take a look at your cardiolipin membrane simulations.
For what I know, the code does not limit in the number of atom names per lipid resname, although the software calls them atomtriads.
I hope this is usefull for you.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
According to MEMBPLUGIN defaults (used in our paper (doi:10.1093/bioinformatics/btu037)) and by analogy, selections for phospholipids and sphingolipids consist on the following 3 atoms:
For sterols such as cholesterol, the selection consists on a polar atom of the polar head (O of the hydroxy group).
Keep in mind that modular topologies like AMBER Lipid14 force field, where a single lipid molecule has several residue names, are not curently supported at least for Area per Lipid computations. You will need to rename your molecules if this is your case.
Wiki: Area per Lipid Tool
Related
Wiki: AreaPerLipid
Last edit: Ismael Rodriguez Espigares 2018-01-18
Do you have any suggestion of selection for more complex lipids such as PMCL?
I have no experience with cardiolipins, but probably taking the atom names for PMCL by doing an analogy among two phospholipids and PMCL could be a good approach. This should result into a set of 6 atom names ("C12 CA1 CB1 C32 CC1 CD2" in CHARMM) for the resname "PMCL" .
C12 = C32 = phospholipid C2
CA1 = CC1 = phospholipid C21
CB1 = CD1 = phospholipid C31
Maybe, you can add extra atoms that are usually in the same plane and in the rigid region of the lipid. The objective here is that the selected atoms define a membrane plane and the bulkiness of the lipid in that plane. Just take a look at your cardiolipin membrane simulations.
For what I know, the code does not limit in the number of atom names per lipid resname, although the software calls them atomtriads.
I hope this is usefull for you.