Hi Brian, Both sqm and mopac are doing a QM charge calculation, which considers the overall electronic environment. The more negatively charged carboxylate oxygen is on the "side" of the molecule that has the positively-charged guanidinium group which seems to cause some of the negative charge to be shifted to that oxygen. Chopping out the guanidinium group results in the carboxylate oxygens being given charges that differ by 0.011. If I eliminate all other ring substituents, which leaves a much truer symmetric environment for the carboxylate oxygens, then sqm computes equal charges for the oxygens (mol2 file attached): 5 C6 23.4850 19.6920 64.8870 C.2 1 BCZ 0.9126 6 O8 22.2660 19.7760 65.0340 O.co2 1 BCZ -0.8493 7 O7 24.2040 20.8120 64.7970 O.co2 1 BCZ -0.8493 I guess the take home is that the asymmetric charges on the carboxylate oxygens is actually the expected behavior. Now, if you feel that the charge transfer is unrealistically high you would have to take that issue up with the folks on the Amber mailing list -- I certainly don't have the expertise to make those kinds of changes! Or if you felt that Chimera should be invoking sqm with something other than sqm's default settings, that is something I could offer control over at least. Let me know if that is the case. --Eric On May 27, 2011, at 8:48 AM, Brian Fochtman wrote:
Hi,
When adding partial charges (AM1-BCC) to small molecules in chimera, we observed that atoms in symmetrical groups are not assigned the same partial charge. Specifically, the two oxygens in a carboxyl group are not assigned the same partial charge. See excerpt from mol2 file below:
carboxyl: 20 C20 23.4850 19.6920 64.8870 C.2 1 BCZ 0.9226 21 O21 22.2660 19.7760 65.0340 O.co2 1 BCZ -0.9253 22 O22 24.2040 20.8120 64.7970 O.co2 1 BCZ -0.7643
There is a significant difference in partial charge between the two carboxyl oxygens (0.161).
This problem was observed in both older (with mopac) and newer (with sqm) versions of chimera.
The molecule is peramivir from the pdb1L7F crystal structure. Similar problems were observed for other molecules that inhibit Neuraminidase for example those 1L__ pdb structures.
This problem is also observed when assigning am1bcc charges to the mol2 file using antechamber/sqm directly.
Best Regards,
Brian Fochtman, Trent Balius, Sudipto Mukherjee
<1l7f.lig.mol2>_______________________________________________ Chimera-users mailing list Chimera-users@cgl.ucsf.edu http://plato.cgl.ucsf.edu/mailman/listinfo/chimera-users