Dear LAMMPS user,
I am trying to generate LAMMPS data file for heptadecane using charmm2lammps tool. The LAMMPS version I am using is 29 Oct 2020 and charmm2lammps v1.9.2 (c)2005-2018. I am using par_all36_lipid.prm and top_all36_lipid.rtf for conversion, as attached to this email.
But, somehow, the generated LAMMPS data showed only one atom type as:
53 atoms
52 bonds
102 angles
144 dihedrals
0 impropers
1 atom types
0 bond types
0 angle types
0 dihedral types
0 improper types
The terminal showed as :
Info: using Heptadecane.pdb instead of Heptadecane.crd
Info: lx not set: will use extremes
Info: ly not set: will use extremes
Info: lz not set: will use extremes
Info: creating PSF index
Info: converting atoms
Info: converting bonds
Warning: bond parameter 1 for [CRL2 CRL2] was not found
Info: converting angles
Warning: angle parameter 1 for [CRL2 CRL2 CRL2] was not found
Info: converting dihedrals
Warning: dihedral parameter 1 for [CRL2 CRL2 CRL2 CRL2] was not found
Info: converting impropers
Info: conversion complete
Would anyone suggest to me please how to solve this? Thank you in advance.
Regards,
Togi
Heptadecane.psf (12.5 KB)
Heptadecane.pdb (4.19 KB)
par_all36_lipid.prm (28.8 KB)
top_all36_lipid.rtf (635 KB)
The problem is that charmm2lammps.pl does not recognize any of your assigned atom types in the .psf files since they are invalid. However you created the .psf file is bogus.
The telltale sign is that you get these warnings:
Warning: bond parameter 1 for [CRL2 CRL2] was not found
Warning: angle parameter 1 for [CRL2 CRL2 CRL2] was not found
Warning: dihedral parameter 1 for [CRL2 CRL2 CRL2 CRL2] was not found
where CRL2 is the last atom type listed in the ATOMS section of the .prm file.
GI-GO,
Axel.
Hi, Axel
Thank you for your response.
Actually, I followed the steps as in https://github.com/lammps/lammps/tree/master/tools/ch2lmp to generate the *.psf file. I re-try the examples on that page and it worked. But, somehow, it is not successful when I try for the molecule I am working on.
Regards,
Togi
The problem seems to be twofold.
- You need to create a valid X-PLOR style .psf file (i.e. with numeric atom types).
- You need to have a compatible CHARMM parameter and topology file
as I already mentioned, your .psf file is bogus: all atom types are listed as type “1”. that is not correct.
also, your parameter file uses the new “stream” format, where there are no more numeric atom types (they are all set to 1 and CHARMm (the code) ignores them). This is not compatible with charmm2lammps.pl which predates this by many, many years. this code was originally conceived for CHARMM version 19. the bundled CHARMM files are in the non-stream format, BTW.
on top of that, your psf file seems to have a completely garbled bond topology (as can be seen from visualizing it with VMD, for example). so even after fixing the atom type assignment it will not produce a working simulation.
axel.
Togi,
There is a way to get the script to work.
Look at the MOST RECENT Version of the par_all36_cgenff.prm file:
!hydrogens
MASS -1 HGA1 1.00800 H ! alphatic proton, CH
MASS -1 HGA2 1.00800 H ! alphatic proton, CH2
MASS -1 HGA3 1.00800 H ! alphatic proton, CH3
MASS -1 HGA4 1.00800 H ! alkene proton; RHC=
.
.
.
Now look at an older version
!hydrogens
MASS 256 HGA1 1.00800 ! alphatic proton, CH
MASS 257 HGA2 1.00800 ! alphatic proton, CH2
MASS 258 HGA3 1.00800 ! alphatic proton, CH3
MASS 259 HGA4 1.00800 ! alkene proton; RHC=
.
.
.
Notice that in the second column the numbers were replaced by “-”. Each atom type was assigned a number in the older version. But this presented a limitation when a new atom type was to be added because a new atom type needed a unique number.
The “-” replacement allowed more flexibility in the addition of new atom types.
If you can find an older version of the forcefield parameter files, and replace the new files with the old files, the script will work.
Regards,
Angelo
Dear Angelo,
Thank you very much.
It worked after I replace with the older version. The older version is in https://github.com/ChayaSt/Charmm36_FF/tree/master/toppar.
Thank you.
Regards,
Togi
Dear LAMMPS users,
Continuing previous email about charmm2lammps tools, unfortunately, I found the inconsistency of dihedral between *.psf file and generated LAMMPS data, as attached to this email.
In *.psf file, there are 144 dihedrals, while 186 dihedrals in LAMMPS data. Anyone can help with this?
Thanks in advance.
Regards,
Togi
1xheptadecane.data (29.2 KB)
1xheptadecane.psf (12.5 KB)
Dear LAMMPS users,
Continuing previous email about charmm2lammps tools, unfortunately, I found the inconsistency of dihedral between *.psf file and generated LAMMPS data, as attached to this email.
In *.psf file, there are 144 dihedrals, while 186 dihedrals in LAMMPS data. Anyone can help with this?
Have you looked at the definition of the Dihedrals more closely and also the corresponding parameters?
for example the dihedral of atom types CTL2 CTL2 CTL2 CTL2 has 4 (four!) sets of parameters with different multiplicity:
CTL2 CTL2 CTL2 CTL2 0.101 2 0.00 ! alkane, 7/08, jbk
CTL2 CTL2 CTL2 CTL2 0.142 3 180.00 ! alkane, 7/08, jbk
CTL2 CTL2 CTL2 CTL2 0.074 4 0.00 ! alkane, 7/08, jbk
CTL2 CTL2 CTL2 CTL2 0.097 5 0.00 ! alkane, 7/08, jbk
and the first four DihedralCoeffs dihedral types are:
1 0.101 2 0 1 # CTL2 CTL2 CTL2 CTL2
2 0.142 3 180 0 # CTL2 CTL2 CTL2 CTL2
3 0.074 4 0 0 # CTL2 CTL2 CTL2 CTL2
4 0.097 5 0 0 # CTL2 CTL2 CTL2 CTL2
now look at Dihedrals 16 to 19:
16 1 5 8 11 14 # CTL2 CTL2 CTL2 CTL2
17 2 5 8 11 14 # CTL2 CTL2 CTL2 CTL2
18 3 5 8 11 14 # CTL2 CTL2 CTL2 CTL2
19 4 5 8 11 14 # CTL2 CTL2 CTL2 CTL2
so it all makes sense there must be more dihedral definitions in LAMMPS than in a .psf file where the multiplicity is implicit.
Axel.
Dear Axel,
I greatly thank you for your response.
Regards,
Togi