GROMACS version: 2022.3
GROMACS modification: Yes/No
Hi there,
I am struggling to run gmx do_dssp. I have tried several options. Here it is:
gmx do_dssp -s sim_eq-3.tpr -f sim_eq-3.xtc -o -sc -ver 4
after selecting MainChain I got:
There are 299 residues in your selected group
dssp cmd=‘/group/davila/SOFTW/COMPCHEM/bin/dssp --output-format dssp ddhzTGP0’
Reading frame 0 time 0.000
Back Off! I just backed up ddhzTGP0 to ./#ddhzTGP0.1#
Error parsing PDB at line 9
When validating _exptl.method: Value ‘PURE PRODUCT OF COMPUTER SIMULATION’ is not in the list of allowed values
If I delete line 9 “EXPDTA PURE PRODUCT OF COMPUTER SIMULATION” and I run
dssp ddhzTGP0
everything run fine with the wished output.
I compiled my own mkdssp version 4.2.0. I tried also other version 4.2.1, 4.1.0, 4.0.0, 2.x (using -ver 2) but I cannot run dssp on my system.
Any help??
This seems to be a recent change with version 4.2.0
I can confirm version 4.0.2 works with -ver 4. You can dowload it from the dssp v4 release page. You need a recent version of gcc to compile it (tested with gcc-11, but version from 8 up should work).
For information, gromacs 2023 has a built-in DSSP utility (gmx dssp) that is less cumbersome to build/use, and faster. You can find the release candidate here.
1 Like
Thanks @ebriand . Which version of libcifpp? I see differences too in this library.
Oh yes, that is probably the root cause of the issue. I’m not entirely sure, but it seems like I have Version 2.0.3 installed - so likely this one.
I tried 2023rc1 and it worked using gmx dssp. Is it reliable as gmx do_dssp? The help says:
“Note that gmx dssp currently is not capable of reproducing the secondary
structure of proteins whose structure is determined by methods other than
X-ray crystallography (structures in .pdb format with incorrect values in the
CRYST1 line) due to the incorrect cell size in such structures.”
That means that trajectory from NMR and/or homology structure is unreliable?
Thanks
Hi,
I tried the version you said, dssp 4.0.2 and libcifpp 2.0.3 - unsuccessfully. New error:
Selected 17: ‘r_1-10_&_MainChain’
There are 10 residues in your selected group
dssp cmd=‘/group/davila/SOFTW/COMPCHEM/bin/dssp --output-format dssp ddh9Ej4z’
Reading frame 0 time 0.000
Back Off! I just backed up ddh9Ej4z to ./#ddh9Ej4z.1#
Resulting mmCIF file is not valid!
Invalid mmCIF file use --verbose option to see errors
Not a known element:
dssp standalone:
dssp ddh9Ej4z --verbose
unrecognized file extension, trying cif
Not cif, trying plain old PDB
CCP4 monomers library not found, CLIBD_MON is not defined
Error validating method: When validating _exptl.method: Value ‘PURE PRODUCT OF COMPUTER SIMULATION’ is not in the list of allowed values
missing mandatory field type_symbol for Category atom_site
Resulting mmCIF file is not valid!
Error validating method: When validating _exptl.method: Value ‘PURE PRODUCT OF COMPUTER SIMULATION’ is not in the list of allowed values
missing mandatory field type_symbol for Category atom_site
Invalid mmCIF file.
Not a known element:
is this something new?
Thanks
I would say it is reliable. It produces the same structure classification as dssp, except for occasional difference at the end of secondary structure tract (e.g. dssp will say that some residue at the end of an helix is classified helix, gmx dssp will say its not - meanwhile the rest of the tract is predicted identically).
The remark is not really about the method - its just that PDB file for non-X-ray crystallography structure can have a CRYST1 line with unit cell 1.0 1.0 1.0 – which will mess up the distance calculation since it relies on an incorrect box size. You can use gmx editconf to change the file format to .gro with a large enough box, and the program will run fine.
I can’t really help you with the DSSP 4 issue, since I cant reproduce the problem. If you want gromacs to stop emitting the EXPDATA line, remove line 314 in src/gromacs/fileio/pdbio.cpp, then rebuild gromacs. Otherwise here is a link to a dssp v2 source I was able to build from. Good luck.
1 Like
Thanks a lot for you kind reply and help. I will use 2023rc1
Best