Parameter Optimization Tutorial

Objective and Overview

>The objective of this tutorial is to demonstrate how to extend the CHARMM additive biological force fields to drug like molecules. The tutorial is based on the CHARMM force fields, parametrization strategies and future/ongoing developments lecture given at the 2006 MMTSB/CTSB Workshop. Reading the powerpoint presentation from the workshop is suggested prior to attempting the tutorial. The molecule for which parameters will be assigned and developed is shown on the right.


Step one of the process is to break the compound into molecular fragments and create the corresponding model compounds. This process yields indole, phenol and the model compound B shown below. The remainder of the tutorial focuses on obtaining and optimizing parameters for compound B. However, the RTF and parameters for the entire drug molecule, along with that for model compound B, are included in the toppar stream file top_mmtsb_example.str. Note that top_mmtsb_example.str is read after the files top_all27_prot_lipid.rtf and par_all27_prot_lipid.prm located in the toppar subdirectory as performed in the input files given below.


The initial charmm script, gen_model_b.inp, generates B and minimizes the structure in 4 different conformations. In addition, the script writes input files for the Gaussian QM package that may be used to perform minimizations and frequency calculations (gauss subdirectory). Several inputs for the optimization of parameters associated with B are included.

water_model_b.inp: Calculates minimum interaction energies and distances of water with B.

model_b_molvib.inp: Calculates the vibrational spectra including performing normal mode assignments. The data in the resulting output are then compared with results from a QM frequency analysis that may be done in CHARMM using the script model_b_molvib_g03.inp.

model_b_surf_all_one.inp and model_b_surf_all_two.inp calculate several potential energy surfaces for rotation around bonds in B. The two files are run consequtively to obtain surfaces that are offset to zero; this could be performed in one file if so desired. The resulting surfaces may be compared to QM surfaces (see *.map files that include the basis set name mp2_631gs).

For the full drug compound, gen_drug.inp, is used to generate and minimize the compound and, model_2_surf_all.inp, calculates two potential energy surfaces for comparison with QM data (see *

Subdirectory gauss: The subdirectory contains a number of input files for the Gaussian program including those created by gen_model_b.inp and gen_drug.inp. See the 00readme file in that directory for more information.

Written by A.D. Mackerell, Jr.