Re: [AMBER] Add molecule to prmtop/inpcrd ?

From: Gustavo Seabra <gustavo.seabra.gmail.com>
Date: Wed, 7 Apr 2021 15:29:17 -0400

Hi all,

Thanks a lot Dan and Matias for their answers. The problem really is more
complicated than one would initially imagine. I believe I've found a way
that seems to work for me, and mostly skipped LEaP. so far, the files are
woring and the simulation is running fine.

For anyone interested, here's how I made it work:

1. Use Antechamber / parmchk2 to create parameters for the ligand. I use
two steps because I want to have a PDB file with Sybyl atom types as well:

$ antechamber -i ligand.pdb -fi pdb -o ligand.sybyl.mol2 -fo mol2 -at sybyl
$ antechamber -i ligand.sybyl.mol2 -fi mol2 -o ligand.gaff2.mol2 -fo mol2
-at gaff2 -c bcc
$ parmchk2 -i ligand.gaff.mol2 -f mol2 -s gaff2 -o ligand.frcmod


2. Use LEaP to create parm7/rst7 files for the ligand alone:

*$ cat leap.in <http://leap.in>*
source leaprc.gaff2
loadamberparams ligand.frcmod
LIG = loadmol2 ligand.gaff2.mol2
saveamberparm LIG ligand.parm7 ligand.rst7
quit


3. Use ParmEd to combine the topologies and restart files. For that, I used
a python script as follows:

*$ cat combine.py*
import parmed as pmd
parm1 = pmd.load_file('ligand.parm7','ligand.rst7')
parm2 = pmd.load_file('protein.parm7','protein.rst7')

# Combine the two files. Note the ligand comes first, which (IMHO) is better
# than having it after all else.The downside is that the new combined parm
# does not have a bounding box, which is needed for the simulation.
comb = parm1 + parm2

# This puts a bounding box in the new parm. We basically
# just copy the bounding box from the first protein system:
comb.box = parm2.get_box()[0]

# Now, simply adding the ligand makes it overlap with waters that were there
# before. This strips only the waters that collide with the ligand:
comb.strip(":LIG(<:1.0&:WAT)")

# Finally, save the files:
comb.save('protein-ligand.parm7')
comb.save('protein-ligand.ncrst',format='ncrst')
comb.save('protein-ligand.pdb'


I looked at the structures in VMD (to load parm7/ncrst), and they seem to
have the correct bonds. The simulation is running fine (so far...)

Thanks a lot!

--
Gustavo Seabra.
On Wed, Apr 7, 2021 at 10:32 AM Daniel Roe <daniel.r.roe.gmail.com> wrote:
> Hi Gustavo,
>
> Welcome to the ordeal that is trying to get LEaP to work with sugars.
> I've recently butted my head up against this myself. I have a
> work-in-progress command in cpptraj called 'prepareforleap' that I've
> been testing that tries to make this process a bit easier. It will
> disconnect any sugars, try to rename them properly, and write out the
> leap input needed to recreate the bonds after you load the prepared
> structure into leap (it also creates leap input to bond disulfides and
> changes CYS to CYX etc). Again, it's still a work in progress so it
> may not work for your system. If you are feeling brave and want to
> give it a shot, email me off-list and I'll give you more details.
>
> -Dan
>
> On Tue, Apr 6, 2021 at 5:27 PM Gustavo Seabra <gustavo.seabra.gmail.com>
> wrote:
> >
> > Hi all,
> >
> > I wonder if there's any way to add a molecule to previously prepared
> > prmtop/inpcrd pair, without using any intermediate file (pdb, etc.)
> >
> > Can parmed do that?
> >
> > The reason I ask:
> > I did a simulation of a reasonably large system, with a membrane-inserted
> > glycosylated protein. Now, I'd like to add a ligand to a specific frame
> > from the trajectory I obtained, and run a new simulation in the presence
> of
> > this ligand.
> >
> > I tried the obvious: use cpptraj to select the frame, write as PDB, add
> the
> > ligand and use LEaP again to build the new system, but the problem is
> that
> > the new PDB saved by cpptraj either does not have CONECT information, or
> > has it for ALL bonds. (The same happens with AMBPDB). If I write with NO
> > CONECT, LEaP then doesn't read the glycam bonds correctly. If write the
> > CONECT cards, LEaP gets confused with some atoms having too many bonds
> > (apparently it may be caused by numbers overlapping in the PDB format)
> >
> > Thanks,
> > --
> > Gustavo Seabra.
> > _______________________________________________
> > 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
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber

Received on Wed Apr 07 2021 - 12:30:03 PDT