[Prev][Next][Index][Thread]
Re: WSN: Binding energy calculations in water solvent
From: Xinjun J. Hou <hou@agouron.com>
We have been studying implicit solvation models implemented in
Macromodel/Batchmin from Clark Still's group to calculate
ligand-protein binding energy.
Our approach is based on energy minimization of ligand-protein complexes
(with GB/SA). Protein is truncated to a small set for faster minimization,
and has shells of increasing constraints. Protein residues near
binding site have full flexibility.
We could find marginal correlation for diversed set of ligands. Better
correlation could be found for same series of compounds. (small variation)
We als found that GB/SA solvation energy is sensitive to conformations
and ligand partial charge assignment. For de novo design, prediction
of bound conformation is also critical. Cancellation of errors plays an
important part in this approach.
We are in the process of studying multiple conformation averaging
with this method. We did study the contribution of ligand
conformational energy in getting better correlation, but it is not
very sucessful yet.
Xinjun
C Xinjun J. Hou hou@agouron.com
C Agouron Pharmaceuticals, Inc. San Diego, CA 92121
C10110000110100101101110011010100111010101101110010010000110111101110101