On Thu, 2014-03-27 at 01:42 +0800, Nitin Sharma wrote:
> Respected sir,
>
> Thanks for prompt reply and yes that solved my problem. Only one more
> query regarding generation of stripped prmtop. As I can generate the
> prmtop files by following two which I can see differs slightly in
> their size. So, which process is the correct approach?
The difference in size stems from a fundamental difference in how
cpptraj and ParmEd "compute" a topology file. cpptraj eliminates the
entries for each of the stripped atoms in the atom property arrays (like
MASS, ATOM_NAME, etc.) and renumbers all of the pointers in the bond,
angle, and dihedral arrays (as well as the exclusion lists). It does
not, however, trim the bond/angle/dihedral type arrays to eliminate the
parameters that are no longer in use. This prevents cpptraj from having
to keep track of those types (if you never touch those arrays, then you
never have to adjust pointers _into_ that array which simplifies the
bookkeeping). I'm not sure if this is still true for Amber14, but Dan
can correct me if I'm wrong.
ParmEd, however, recomputes the _entire_ topology file and adds types
into the parameter file as they're used. This approach eliminates all
unused parameter types and leads, in general, to a smaller topology file
than those written by cpptraj. Both are equally valid topology files,
however, and result in completely equivalent energies and forces. They
look completely different, though, as the pointer and parameter ordering
is completely different (but again, equivalent).
>
> cpptraj complex_solvated.prmtop strip2merge_2 > strip2merge_2.out
> #strip2merge2
> trajin 5md.mdcrd 2000 last
> strip :1-18,168-182,Na+,Cl- outprefix stripped_2NL9_046
> trajout 2NL9_046_strip.mdcrd netcdf nobox
>
> OR
>
> parmed.py complex_solvated.prmtop strip_Parm_2
> #strip_Param_2
> strip :1-18,168-182,Na+,Cl- outprefix stripped_2NL9_046
"outprefix" is not a recognized keyword for ParmEd (in cpptraj it is
only used to trigger a prmtop dump, which doesn't apply in ParmEd). In
the next version of ParmEd this will be a fatal error by default.
> parmout /hpctmp/a0068362/MCL1_combined_test/2NL9_046_strip_prmtop.prmtop
>
> Moreover, as prmtop file is required by cpptraj so do I need to load
> both prmtop files or have to merge prmtop files as well. If I have to
> merge them then how can I do that ?
I'm not sure what you mean by "merge" them. You can use either the
topology file written by cpptraj or the one written by ParmEd -- they
should both work exactly the same for any program. Since you're already
using cpptraj, though, you can easily eliminate ParmEd from your
workflow here.
HTH,
Jason
--
Jason M. Swails
BioMaPS,
Rutgers University
Postdoctoral Researcher
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Mar 26 2014 - 11:30:04 PDT
