Re: Protein-ligand simulation

From: James Starlight (jmsstarlight_at_gmail.com)
Date: Thu Jun 27 2013 - 00:16:25 CDT

Using below script

package require psfgen
resetpsf
topology top_all36_na.rtf
segment A {
pdb input.pdb
first NONE
last NONE
}
coordpdb input.pdb A

guesscoord
patch CY35 A:1 A:1
regenerate angles dihedrals

I've obtained

psfgen) Info: skipping conformation C3'-O3'-P-O5' at beginning of segment.
psfgen) Info: skipping conformation C4'-C3'-O3'-P at beginning of segment.
psfgen) Info: skipping conformation O3'-O5'-P-O2P at beginning of segment.
psfgen) Info: skipping conformation O3'-O5'-P-O1P at beginning of segment.
psfgen) Info: skipping conformation O3'-P-O5'-C5' at beginning of segment.
segment complete.
psfgen) reading coordinates from pdb file input.pdb for segment A
psfgen) Info: guessing coordinates for 0 atoms (0 non-hydrogen)
psfgen) applying patch CY35 to 2 residues
A:1 A:1
psfgen) regenerating all angles
psfgen) regenerating all dihedrals
psfgen) Info: writing psf file ligand.psf
psfgen) total of 34 atoms
psfgen) total of 37 bonds
psfgen) total of 66 angles
psfgen) total of 99 dihedrals
psfgen) total of 3 impropers
psfgen) total of 0 cross-terms
psfgen) Info: psf file complete.
psfgen) Info: writing pdb file ligand.pdb
psfgen) Info: pdb file complete.

Here you can see that dihedrals and angles for new bond were also included
in the topology new bond between O3 and P looks strange :)

     37 !NBOND: bonds
       1 2 1 3 1 4 4 5
       5 8 5 6 5 7 8 10
       8 32 8 9 10 11 11 13
      11 28 11 12 13 14 13 26
      14 18 14 24 15 16 15 17
      19 15 19 20 19 18 20 21
      20 22 22 24 22 23 24 25
      25 26 26 27 28 32 28 30
      28 29 30 31 32 34 32 33
      34
1 # new bond
between P and O

So I still not quite sure about correct application of the patch. Should I
explicitly define atoms which should be patched in the psfgen script (e,g
patch CY35 A:1:1 A:1:34) or separate P group in the individual segment ?

James

2013/6/27 JC Gumbart <gumbart_at_ks.uiuc.edu>

> psfgen autogenerates by default when building segments (but not patches!).
> The closest equivalent would be to add near the end of the script before
> writing out the psf:
>
> regenerate angles dihedrals
>
>
> On Jun 26, 2013, at 3:07 PM, Kenno Vanommeslaeghe wrote:
>
> > Hi James,
> >
> > On 06/26/2013 01:56 AM, James Starlight wrote:
> >> By the way what maximum penalty value can be refined by means of VMD's
> >> topology plugin ?
> >
> > I'm sorry, I don't understand what you mean by this. I don't know "VMD's
> topology plugin", but I doubt it can be used to obtain more reliable values
> for penalties that are high. Please do have a look at my paper on how the
> penalties are calculated.
> >
> >> Here you can see 2 outputs with and without applied patch. As You can
> see
> >> the only defference is 1 extra bond when patch was used without any
> >> additions in dihedral terms. Does it correct in case of cyclic
> nucleotides?
> >
> > Nope, the numbers after patching are not correct. AUTOGEN ANGLE DIHE is
> a CHARMM command. It looks like your problem is that psfgen doesn't
> recognize it, in which case you need to find the psfgen command that is
> equivalent to AUTOGEN ANGLE DIHE .
> >
> > Cheers,
> >
> > Kenno.
>
>
>

This archive was generated by hypermail 2.1.6 : Wed Dec 31 2014 - 23:21:21 CST