I've a problem that has been stumping me for days, I have a peptide structure that I pulled from a MD trajectory. I'm unable to get any Rosetta modules like relax, or fixbb to work with this structure, the jobs always start off fine but as soon as they hit a THR residue of which there are four I get the following:
core.chemical.ResidueType: [ ERROR ] atom name : CZH not available in residue THR
I see nothing wrong in my PDB file, here's the first THR residue:
ATOM 26 N THR A 3 24.927 7.409 -5.735 1.00 0.00 N
ATOM 27 CA THR A 3 25.090 7.952 -7.051 1.00 0.00 C
ATOM 28 C THR A 3 24.503 9.327 -7.033 1.00 0.00 C
ATOM 29 O THR A 3 23.949 9.680 -5.983 1.00 0.00 O
ATOM 30 CB THR A 3 24.348 7.087 -8.137 1.00 0.00 C
ATOM 31 OG1 THR A 3 24.814 7.426 -9.443 1.00 0.00 O
ATOM 32 CG2 THR A 3 22.862 7.269 -8.173 1.00 0.00 C
ATOM 33 H THR A 3 24.285 7.894 -5.146 1.00 0.00 H
ATOM 34 HA THR A 3 26.152 8.057 -7.217 1.00 0.00 H
ATOM 35 HB THR A 3 24.642 6.052 -7.861 1.00 0.00 H
ATOM 36 HG1 THR A 3 25.756 7.256 -9.522 1.00 0.00 H
ATOM 37 1HG2 THR A 3 22.576 8.326 -8.359 1.00 0.00 H
ATOM 38 2HG2 THR A 3 22.329 6.571 -8.854 1.00 0.00 H
ATOM 39 3HG2 THR A 3 22.452 6.954 -7.189 1.00 0.00 H
I've been working primarily with rosetta.binary.mac.release-223 r223 2019.22+release.d8f9b4a90a8, but I get the same behavior running on a linux cluster as well. A related peptide to the one I'm working with with 5 THR's runs fine with everything I've thrown at it. Any suggestions would be greatly appreciated.
Looking for CZH indicates that Rosetta seems to be wanting to type one of your THR residues as an "hydrogen bonding surrogate" type. I'm not entirely sure why that might be. The THR residue you show looks to be formulated in a way which Rosetta should like.
Are the other THR residues identical to this block (except for coordinates and residue numbering)? There's no "extra" atoms in any of your THR?
What's the PDB header like. You say it's from MD, so I expect it's minimal, but are there any directives relating to THR residues? I'm thinking in particular of records which indicate a THR might be bonded to another residue (or have extra bonds within itself.)
If you're willing, it might help to post a sample PDB of a structure with the issue here, and there might be something we can glean from the full PDB structure. (Or someone might be able to run things locally and get a better sense of what's going on.)
I edited out all of the header info in one of my attempts to solve the problem. I've attached the pdb file in question, all of the THR entries are set up identically. In this version I'd hand-edited the THR's to match what ROSETTA seemed to be expecting, despite that I'm still getting the same error as I was seeing with the original un-edited file. I hope that some fresh eyes can figure out what is going on.
Interestingly, it seems like it's *not* the THR residues which are the issue. Instead it's the initial GLY residue, particularly the acetyl group atom nomenclature. The hydrogen bonding surrogate patches use the atom names CY/OY/HY1/HY2/HY3, so when Rosetta reads in your PDB, it thinks that the initial residue is a hbs type. It then tries to find the partner hbs residue at position 3. But in your protein it isn't an hbs type, but rather just a plain THR residue, leading to the issue you see.
If you want to have an acetylated N-terminus, you should use the atom names CO/OP1/CP2/1HP2/2HP2/3HP2 if you include the atoms as part of the initial residue itself, or use an ACE residue (with C/O/CH3/1HH1/2HH3/3HH3 atoms) as a "residue 0".
Sure enough that solved the problem, I would never have tracked that down on my own.