Hi Jun Feng,
This is strange. It should match perfectly. I am summarizing in the table
below to make it clearer.
Charmm AMBER
Bond 149.5654 149.5654
Angle 463.7249 463.7249
Dihe 1517.1368 1517.7891
UB 23.8425 23.8425
IMP 18.5436 18.5436
CMAP 68.836 68.8362
1-4NB 391.8176
VDW -1031.4918
VDWsum -639.6742 (-639.6742)
1-4EEL 6625.6748
EELEC -10081.5617
EELECsum -3455.8869 (-3455.8869)
So you are right the dihedral term does not match. I can think of two
possibilities here. One is that there are some dihedrals in your system
that are very planar - it might be worth running your system for a few ps
to let things equilibrate and then take a snapshot from there. That will
rule out if it is coming from something where AMBER and CHARMM treat
dihedrals at the very fringe slightly differently.
The other possibility is that there are dihedral terms missing. You should
be able to run the prmtop through ptraj (rdparm) and have it print all of
the dihedrals it finds in the prmtop. I think there is a similar way to do
this in Charmm and then one can compare and see if any dihedrals are
missing or different. Have you tried this on something smaller such as
alanine dipeptide? It is strange that it only happens with C36. I believe
at the time Chamber was developed we only had c22 to test. One possibility
could be related to systems with rings - for example cyclopentane - should
not include dihedrals in the ring because there are already 1-3
interactions in the ring.
E.g.
1
/ \
/ \
2 5
| |
3 ---- 4
Dihedral 1-5-4-3 duplicates the angle 1-2-3 so should be excluded. It is
possible this is being treated differently in AMBER and Charmm in the case
of C36. - Just a guess. It will need some debugging to figure out what is
going on here.
All the best
Ross
On 8/30/13 2:15 PM, "Jun Feng" <jufeng.mail.wvu.edu> wrote:
>Dear CHAMBER developer:
>I want to run CHARMM FF in AMBER. My workflow is:
>1. created psf and pdb in CHARMM
>2. use CHAMBER to generate prmtop and inpcrd
>3. run one step energy calculation in CHARMM and sander, compare energy,
>cutoff in both CHARMM and AMBER set to 999
>Result:
>All energy terms in CHARMM FF c22 are the same. All energy terms except
>Dihedral are the same in c36. Dihedral energy is slightly off in c36 no
>matter turning on or off CMAP. I tested on several proteins and here is
>the
>result of DHFR in vacumm in c36:
>
>#AMBER
> NSTEP = 0 TIME(PS) = 0.000 TEMP(K) = 300.85 PRESS =
>0.0
> Etot = 377.0066 EKtot = 2230.2661 EPtot =
>-1853.2594
> BOND = 149.5654 ANGLE = 463.7249 DIHED =
> 1517.7891
> UB = 23.8425 IMP = 18.5436 CMAP =
> 68.8362
> 1-4 NB = 391.8176 1-4 EEL = 6625.6748 VDWAALS =
>-1031.4918
> EELEC = -10081.5617 EHBOND = 0.0000 RESTRAINT =
>0.0000
>
>#CHARMM
>ENER ENR: Eval# ENERgy Delta-E GRMS
>ENER INTERN: BONDs ANGLes UREY-b DIHEdrals
> IMPRopers
>ENER CROSS: CMAPs PMF1D PMF2D PRIMO
>ENER EXTERN: VDWaals ELEC HBONds ASP
>USER
> ---------- --------- --------- --------- ---------
> ---------
>ENER> 0 -1853.91175 0.00000 2.12417
>ENER INTERN> 149.56538 463.72494 23.84247 1517.13684
>18.54358
>ENER CROSS> 68.83613 0.00000 0.00000 0.00000
>ENER EXTERN> -639.67419 -3455.88690 0.00000 0.00000
> 0.00000
>
>Can anybody give me some hint or point out where I did wrong?
>Thanks!
>
>Jun Feng
>_______________________________________________
>AMBER mailing list
>AMBER.ambermd.org
>http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Aug 30 2013 - 15:30:03 PDT
