Dear Dr. Ross Walker:
Can you give me a specific example of how to create ptraj.in file to
explicitly write out all the dihedrals? How to compare with CHARMM? I am
not familiar with AMBER at all. For CHARMM, I can print out every dihedral
angle and energy.
Here is what I have done to dig into the puzzle in c36 FF:
1. I compared the force on every atom and found the force on backbone C,N
and CA (not O) are different between CHARMM and AMBER. Since all the other
energy terms are the safe except dihedral, I would guess the dihedral
energy on atoms involving C, N and CA are different.
2. I looked at the chamber log file, and found out the difference in c22
and c36. First, the number of dihe does not match. Can you explain? why
difference in dihe without hydrogen?
for c22:
Number of dihedrals found: 6701 (in CHARMM PSF)
Created prmtop summary
Number of dihedrals with hydrogen: 4651
Number of dihedrals without hydrogen: 2349
for c36 (same dihe info w/ or w/o cmap):
Number of dihedrals found: 6701 (in CHARMM PSF)
Created prmtop summary
Number of dihedrals with hydrogen: 4651
Number of dihedrals without hydrogen: 3293
Thanks!
Jun Feng
On Fri, Aug 30, 2013 at 6:03 PM, Ross Walker <ross.rosswalker.co.uk> wrote:
> 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
>
--
Jun Feng
Postdoc Fellow
Blake Mertz Research Group
Department of Chemistry
West Virginia University
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Sep 03 2013 - 15:00:03 PDT
