Reading FE from 2PAH

[Anonymous]

Hey Guys,

We are trying to load a cleaned PDB, 2PAH into Rosetta. The Iron is in the active site, and seems to be readable and on in Rosetta.

Basically, we get this:

core.conformation.Conformation: [ WARNING ] missing heavyatom: FE on residue FE 330
core.conformation.Conformation: [ WARNING ] missing heavyatom: FE on residue FE 330
core.conformation.Conformation: [ WARNING ] missing heavyatom: FE on residue FE 330

ERROR: too many tries in fill_missing_atoms!
ERROR:: Exit from: src/core/conformation/Conformation.cc line: 2664

It tries many times and fails. Why is it not recognizing the FE? Is there something wrong with the FE lines in the PDB, they seem to be formatted correctly?

[Anonymous]

Spacing sometimes matter. Rosetta is expecting an atom name “FE..” (where “.” are spaces). It looks like the atom names are correctly spaced in the from-PDB 2PAH.pdb, but I don’t know if the cleaning process you used might have disrupted that. My local copy of Rosetta seems to have no problems reading the from-PDB version of 2PAH (irons included).

[Anonymous]

Thanks Rocco. Yes, it was my cleaning. I had lined 2 letter atom names up at the same space as the one letter atom names. Just like the rest of the PDB, even the format’s own examples! When I checked the difference, I guess we loaded the cleaned version by accident for comparison.

[Anonymous]

If I change the spacing for all two letter atom names to for example “CA..” (Like the PDBformat guidelines say) (Alignment of one-letter atom name such as C starts at column 14, while two-letter atom name such as FE starts at column 13.)

Rosetta can’t read each residue. Is the spacing FE.. Special for FE or just HETATM, or just anything but CA? Are there any generalities I could make?

[Anonymous]

Never mind, found it. For two letter atom names, elements start at 13, Everything else at 14. Guess I’ll just not strip the string!

[Anonymous]

To be technical, the spacing of the atom names is whatever is specified in the residue’s params file (either by -extra_res_fa or in the database). For the most part it matches how the PDB does it, which appears to be the above-mentioned 13-for-two-letters, 14-for-one letter.

Striping strings of whitespace of a PDB file (or even assuming that whitespace separates different elements) is generally not a good idea. The PDB format is very much a Fortran-derived position-sensitive fixed-width format. Trying to use modern text parsing on it tends to lead to heartache.

[Author]

Posts