OPLS-AA/M ff parameters for Alanine - inconsistent
Neenz231 opened this issue · 2 comments
I believe this to be a bug with LigParGen
Issue Information
Software name & Version : LigParGen server from website
Method: I am a first time user of this server, so tested alanine to check if generated partial charges for the residue (after redistributing charges) are consistent with current aminoacid.rtp file for OPLS-AA/M
Observation
Hi,
did you check the dihedral parameters as well? The parameters that it gave are not consistent with OPLS-AA
Charges
With respect to the partial charges returned by LibParGen, it uses the "CM1A" or "CM1A-LBCC" which uses some kind of semi-empirical quantum calculation to compute them. (Forgive me for being vague. I am not an expert at running QM calculations.) I could be wrong, but I presume this was done because the results of running simulations with these newly calculated partial charges will be more accurate than using the old partial charges which are listed in the template/topology files distributed with OPLSAA.
Anyway, the fact that the two results do not agree is not surprising (although perhaps one might be surprised or alarmed to learn that the difference between the charges assigned by these two methods is so large: almost 0.5 in the first example, "N".).
Dihedral parameters:
It's very difficult to very that the dihedral parameters it assigned are correct because LibParGen obfuscates the atom types used for force field lookup. In issues #46 and #41, I suggested including these atom type names in the files generated by LibParGen. If you agree with this suggestion, please post a supporting followup comment to issues #46.
Cheers
-Andrew