GROMACS version: 5.1.2 (also happens in Gromacs 2019)
GROMACS modification: No
I’m trying to simulate a polypeptide with neutral N terminal (NH2) with the OPLS-AA/M force field. When I use the pdb2gmx command to create the gro and top files, I get the following error:
Program gmx pdb2gmx, VERSION 5.1.2
Source code file: /build/gromacs-z6bPBg/gromacs-5.1.2/src/gromacs/gmxpreprocess/toputil.c, line: 477
Fatal error:
tpA = 53191, i= 0 in print_atoms
For more information and tips for troubleshooting, please check the GROMACS
website at “insert gromacs error home page here”
This error does not occur if I choose the NH3+ capping. Only any other capping throws this error. It appears to be an implementation error since I have made no manual change to the FF files (see here: https://mailman-1.sys.kth.se/pipermail/gromacs.org_gmx-users/2016-April/104769.html) but not quite sure. I was wondering if anyone has come across this before or might know something I could do.
I only need one simulation with this capping type so I am trying to avoid downloading and learning how to properly run this specific simulation (which has more complicated steps after simply choosing an N capping) in a totally different software package.
The force field distribution has missing atom types that you need. The oplsm_900 and oplsm_912 atom types do not appear in atomtypes.atp (which triggers this error) nor are they in ffnonbonded.itp so even if you got a topology, it would fail at grompp due to the lack of parameters.
The OPLSAA/M files have nothing to do with CHARMM, so there should be no connection. At least for CHARMM36, CHARMM should support neutral termini without issue because all of those files are automatically converted from the CHARMM files and validated, so there should be no possibility of a missing atom type.
Sorry I worded my response in an unclear way. I do not experience this issue with the CHARMM36 force field in Gromacs.
The link I downloaded the force field files from (http://zarbi.chem.yale.edu/oplsaam.html) shows “CHARMM and Gromacs formatted parameter files for the OPLS-AA/M force field for proteins”. The first two links show file formats for the CHARMM package. My second question was asking if I could assume that since the atom types I need are missing in the Gromacs files would they would also be missing in the CHARMM formatted files.
That I don’t know, having never tried to use them. But I know users have encountered this problem before - someone should provide the developers some feedback that the force field files (at least for GROMACS) do not work in all cases.
@BrianAndrews Choosing NH2 as terminus using ter command during pdb2gmx never worked for me, instead I already had NH2 in my PDB file and then OPLS recognized it as a residue
Thanks for the response. Could you provide a couple more details about your pdb file or the steps you take? Did you make the NH2 neutral terminal a separate residue from the N terminal amino acid? How did you generate a topology file without pdb2gmx (which automatically attaches an third hydrogen without the ter flag)?
@BrianAndrews Sorry I was referring to C-terminus amide capping (NH2), but your issue is with N-terminus. With OPLS-AA ff, for neutral N-terminus NH2, I had no problem selecting NH2 using pdb2gmx -ter -ignh. The N-terminal tdb already had parameters for neutral NH2 as shown below:
Thanks for the response @Neena. I do not experience this problem in OPLS-AA or OPLS-AA/L which is what the file excerpt you display is from. Atom types in OPLS-AA/M are oplsm_x for updated parameters which the new termini databases show in the files I have linked previously in the thread.
@BrianAndrews in reference to your question about the CHARMM files, those don’t even offer the neutral N-terminus patch (NNEU) so the same problem won’t persist, but one wouldn’t be able to generate a neutral terminus in this case. It appears that the authors added the GROMACS [NH2] terminus to the aminoacids.n.tdb file for compliance with existing GROMACS files but neglected to add all the atom types into the force field. Please contact the OPLS-AA/M authors for a solution.
