Previous Thread
Next Thread
Print Thread
Page 1 of 2 1 2
about adding new force field parameters
#1011 03/03/04 08:12 AM
Joined: Feb 2004
Posts: 30
Forum Member
OP Offline
Forum Member
Joined: Feb 2004
Posts: 30
Hi, All:
I have some questions needing your help. I am eager for your response! thanks a lot in advance!
A) The treatment of atom partial charges. Your tutorial said that just move the hydrogen charge into that of the carbon of model compound if the hydrogen was substituted for a model. What should I do if not just a hydrogen was substituted, for example the schematic drawing: Can I just modify the partial charges of C1 and N only? and how?










B) I still can’t understand the role of “IC” of the topology file of CHARMM force field well, are they only useful for guessing hydrogen atoms? Should I build complete “IC” list when I try to build one new topology (including the impropers and dihedrals, by the way, the four atoms forming improper must be in fixed order?), I can list them for heavy atoms just based on “x-ray” structure, but how to list them for hydrogen atoms?
C) How can I get dihedral parameters with GAUSSIAN, is it directly or not? In gas phase or condensed phase? Would you please send me some of typical gaussian input file?
D) In the parameter file “par_all22_sugar.inp”, every dihedral has three parameter sets, what’s the meaning of them? and which one will be used when do simulations?

Re: about adding new force field parameters
shouqinlv #1012 03/04/04 04:46 PM
Joined: Oct 2003
Posts: 782
A
Forum Member
Offline
Forum Member
A
Joined: Oct 2003
Posts: 782
First, obtain the Sanibel 2003 conference powerpoint presentation
from my web page. That will give you an overview of the steps
in developing parameters. Second, read rtf.doc and parmfile.doc.

http://www.pharmacy.umaryland.edu/faculty/amackere/param/force_field_dev.htm

A) Charmm charges are modular, such that different moieties/fragments are an
integer charge. When combining individual fragments these integer charges
should be maintained. While each situation differs, the general rule
is to put the non-integer charge remaining from the atoms being
removed on the carbon from which they were deleted. In the end, if
you're not sure about the charges you assigned, calculate interactions of
the model compound with water at the HF/6-31G* level and see
how well charmm reproduces them (see the powerpoint presentation
and the psc tutorial for examples).

B) read intcor.doc. the IC table is used primarily to build atoms and
for analysis . the contents of the IC table will not effect the energy
calculation. note that after you create the topology for a new
residue and then generate it in charmm you can use the
"ic generate" command to create an ic table that, with some
reformating, can be placed into the topology file.

C) dihedral parameters are optimized based on the conformational
energies. use gaussian to obtain the QM change in energy as a
function of a selected dihedral. then use charmm to calculate
the same energy surface and adjust the appropriate dihedral
parameters in charmm so the charmm energy surface reproduces
the QM surface.

D) multiple dihedrals allow for a fourier series of cosine terms to be
applied to a single dihedral. this allows for much greater flexibility
in reproducing QM conformational energies. see parmfile.doc and
rtf.doc.

alex





School of Pharmacy
University of Maryland
20 Penn Street
Baltimore, MD, 21201
more questions about building parameter sets!
alex #1013 03/09/04 06:59 AM
Joined: Feb 2004
Posts: 30
Forum Member
OP Offline
Forum Member
Joined: Feb 2004
Posts: 30
Thanks for your answers! After reading more documents, more questions needing your help again:
A) I intended to use GAUSSIAN98 program for the ab initio calculations, but even after reading its user’s guide several times, I still don’t know how to calculate the interaction between model compound and water molecule by using GAUSSIAN98, would you please give me some clue?
B) I am still reading the documents about CHARMM program and not familiar with it enough. So I want to know whether the CHARMM program can carry out the parameter optimization automatically after being given the target data?
Thanks again!
Shouqin Lv

Re: more questions about building parameter sets!
shouqinlv #1014 03/11/04 10:34 PM
Joined: Oct 2003
Posts: 782
A
Forum Member
Offline
Forum Member
A
Joined: Oct 2003
Posts: 782
Man, get some books and learn the basics! QM interactions are determined
by calculating the energy of the complex and subtracting the sum of the
energies of the monomers (with other considerations.....). See my publication; Huang and MacKerell
JPC in 2002/3 for methods to calculate interaction energies. NOTE that
for compatibility with CHARMM22/27 use HF/6-31g* level of theory.

No, charmm does not automatically adjust parameters. you have
to calculate the quantity, compare it to QM or experiment and then
adjust the parameters.

alex


School of Pharmacy
University of Maryland
20 Penn Street
Baltimore, MD, 21201
Re: more questions about building parameter sets!
alex #1015 04/01/04 01:27 PM
Joined: Feb 2004
Posts: 30
Forum Member
OP Offline
Forum Member
Joined: Feb 2004
Posts: 30
Dear Sir:
More questions about optimization of force field parameter sets,
1) Following your suggestions, in order to optimize the partial atom charge, I optimized a molecule structure first by GAUSSIAN98 in order to calculate the interaction between it and water molecule. My question is: which kind of initial structure should be used for the optimization? Must I give Z-matrix definition of the molecule in order to show symmetry, or is it also OK if I give cartesian coordinates definition which are gotten from add-hydrogened PDB file? I did optimization of a molecule with the structure definition by Cartesian coordinates, and found that the partial atom charge did not show necessary symmetry, for example, the partial atom charges of the three hydrogen atoms of model CH3 are different.
2) When I did the interaction energy optimization between a biomolecule and a water molecule by giving relative position and a variable distance through GAUSSIAN98. Sometimes the resulted interaction energy is positive, this phenomena did not change even though I changed the relative position of these two molecules. Did you meet this phenomena before, and what’s wrong with this?
Regards!
Shouqin Lv

Re: more questions about building parameter sets!
shouqinlv #1016 04/01/04 09:10 PM
Joined: Oct 2003
Posts: 782
A
Forum Member
Offline
Forum Member
A
Joined: Oct 2003
Posts: 782

you can use cartesian coordinates for the QM optimizations.

since you are determining the charges based on the reproduction
of hf/6-31g* QM interactions with water, you assign the symmetry
to the charges. yes, all the hydrogens on a methyl group will
be equal.


School of Pharmacy
University of Maryland
20 Penn Street
Baltimore, MD, 21201
Re: more questions about building parameter sets!
alex #1017 04/01/04 10:19 PM
Joined: Sep 2003
Posts: 250
Forum Member
Offline
Forum Member
Joined: Sep 2003
Posts: 250
Just a humble viewpoint: Using HF does not include the electron-correlation vs energy-differences in going from a+b to ab, and therefore perturbation-, or coupled cluster theories and to some extent DFT should be used when estimating such energies. Further, inclusion of zero-point energy corrections and solvent effects changes the energy hypersurface as well.

Re: more questions about building parameter sets!
alex #1018 04/05/04 08:24 AM
Joined: Feb 2004
Posts: 30
Forum Member
OP Offline
Forum Member
Joined: Feb 2004
Posts: 30
I am not familar with the CHARMM program and did not buy it until now, so I want to know: except the program CHARMM, is there any other program can do the parameter sets optimization?

Re: more questions about building parameter sets!
shouqinlv #1019 04/05/04 10:17 PM
Joined: Oct 2003
Posts: 782
A
Forum Member
Offline
Forum Member
A
Joined: Oct 2003
Posts: 782

Sorry, there are not other programs that do parameter optimization,
although charmm does have all the utilities to do it.


School of Pharmacy
University of Maryland
20 Penn Street
Baltimore, MD, 21201
Re: more questions about building parameter sets!
jb007 #1020 04/05/04 10:20 PM
Joined: Oct 2003
Posts: 782
A
Forum Member
Offline
Forum Member
A
Joined: Oct 2003
Posts: 782

hf/6-31g* should be used to maintain consistency of the charges
with the remainder of the force field. you are correct about all the
terms that are being omitted; however, the approach you suggest
would give you good gas phase interaction energies whereas
charmm is optimized to give you condensed phase properties.

alex


School of Pharmacy
University of Maryland
20 Penn Street
Baltimore, MD, 21201
Page 1 of 2 1 2

Moderated by  alex, lennart, rmv 

Link Copied to Clipboard
Powered by UBB.threads™ PHP Forum Software 7.7.4
(Release build 20200307)
Responsive Width:

PHP: 5.6.33-0+deb8u1 Page Time: 0.011s Queries: 34 (0.003s) Memory: 0.9900 MB (Peak: 1.1213 MB) Data Comp: Off Server Time: 2020-09-29 14:03:59 UTC
Valid HTML 5 and Valid CSS