Parametrizing a novel residue

From: Q=C1=D2=D8=D1_=FB=C1=CC=C1=C5=D7=C1?= (skorovesna_at_inbox.ru)
Date: Fri Sep 18 2009 - 10:25:18 CDT

Dear NAMD users,

I am trying to make a topology of a new residue - ALA covalently bound
to SER-Ogamma by an amid bond (to study acil transfer reactions).
As I understood from the manual I need to specify the following things.

in topology
1/ charges for all atoms - by a float number
2/ single and double bonds - by specifying pairs of atoms
3/ impropers - by specifying 4 atoms to have an improper for

in parameter file
4/ parameters for all new bonds - equilibrium length and spring constant

I found several methods to calculate the charges (see below). I tested
them on a serine residue and got the following results:
> RESI SER 0.00
> ! genuine! VEGA ZZ ! !
> ATOM ! CHARMM ! Ammp-Mom Gasteiger ! MOPAC ! GAMESS+RESP (ESP-A1)
> ATOM N NH1 -0.47 -0.64 -0.03 0.08 -0.72
> ATOM HN H 0.31 0.34 0.28 0.76 0.48
> ATOM CA CT1 0.07 0.07 0.23 -0.29 -0.29
> ATOM HA HB 0.09 0.18 0.07 0.16 0.18
> ATOM CB CT2 0.05 0.05 0.08 0.08 0.31
> ATOM HB1 HA 0.09 0.18 0.06 0.04 0.04
> ATOM HB2 HA 0.09 0.17 0.06 0.10 0.04
> ATOM OG OH1 -0.66 -0.61 -0.39 -0.23 -0.66
> ATOM HG1 H 0.43 0.40 0.21 0.24 0.40
> ATOM C C 0.51 0.25 0.30 0.34 0.84
> ATOM O O -0.51 -0.40 -0.26 -0.25 -0.55
>
> Structure
>
> H O H H O H
> | || | | || |
> H-C-C------N--CA-C-------O-C-H
> | | |
> H HB1-CB-HB2 H
> |
> OG
> |
> HG1
>
It could be seen that all methods have a deviation from the genuine
CHARMM force field with Ammp-Mom method showing the best performance.
MOPAC and GAMESS+RESP constantly show negative charge on CA atom.
The next problem that I have spotted is that in CHARMM all atoms are
grouped (GROUP) to have a local zero charge. Unfortunately this is not
possible if using calculated values.

So, my first question is - which method should use and should I try to
make a local groups of zero charge by correcting a little bit the
calculated values?

My second question is - if I have to specify impropers as
IMPR N -C CA HN C CA +N O
IMPR NL CL CLA HLN1 CL CLA NL OL
IMPR CL OG OL CLA
how does the program knows how exactly to restrain the angle?

And my third question is how to calculate bond parameters? I`ve found
the way to calculate it using VEGA, but the manual says that the results
could be rough.

Thank you in advance.
da_shal

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:51:29 CST