AMBER Archive (2005)

Subject: AMBER: Question about partial charge for ligand molecules

From: Woojin Lee (lucius.e_at_gmail.com)
Date: Wed Dec 28 2005 - 20:37:47 CST


Hi folks,

   I generated ligand molecules with MMFFs94 f.f. in Macromodel for binding
affinity MD simulation with Amber, as there is
No crystal structure for lignad-protein complex.
   For consistency, we need to RESP charge. From the structures, we
performed HF 6-31G* single point calculations
with Gamess program to get RESP charges "without geometry optimization" by
the same level of theory and basis sets.

   We examined the am1-bcc by divcon and Antechamber, but we did not find a
way to assign equal charges to symmetric
atoms. So we decided to get partial charge for ligands by RESP protocol.

   What I worry about is that the structure used in single point energy is
not the one from HF-optimization.

what I expect is:
1. MM minimized structures have reasonable geometries.
2. Even we perform HF optimzation from MM-minimized one, it will not alter
the geometry much.
3. Even internal geometry changed by some extent, charge will not change
much.

In addition, the non-parallel Gamess took too much time to run a
optimization job for current ligands( ca. 60 atoms and more...)
(our cluster: Xeon 2.8 GHz RH9)

   Q1) Is there any input keyword or method to assign eual charges for
symmetric atoms by am1-bcc protocol?
   Q2) Is the procedure of ESP generation not from HF-optimized geom, but
from MM-minimized geom can be valid?

Thanks in advance

-------------------------------------
Department of Chemistry
Seoul National University

Woojin Lee

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu