Set b-factor for PDB file

[xiki-tempula]

Expected behaviour

I have a pdb file which I want to set b-factors to the atoms.

Actual behaviour

The b-factor field in the pdb file is not changed

Code to reproduce the behaviour

from MDAnalysis import Universe
from six import StringIO

pdb = 'ATOM      1  N   MET S   1      85.260  60.210  43.770  1.00  0.00      SYST N'
u = Universe(StringIO(pdb), format='pdb')
u.atoms.bfactors = 0.5
u.atoms.write('test.pdb')

The content of test.pdb:

HEADER
TITLE MDANALYSIS FRAME 0: Created by PDBWriter
CRYST1 0.000 0.000 0.000 0.00 0.00 0.00 P 1 1
ATOM 1 N MET S 1 85.260 60.210 43.770 1.00 0.00 SYST N
END

Currently version of MDAnalysis:

py35, mda0.16.1

[kain88-de]

See #1092 . The problem is that the bfactors aren't in the original topology and this needs to be updated not the atomgroup you selected there.

We are working on it. There should also be a PR from @richardjgowers that fixes this. Some comments on the UI/API would be nice.

[orbeckst]

I am going to re-open this one, even though it is a duplicate of #1092 because

1. it is a common use case and anyone being bitten by the new topology system should be able to find this as an open issue
2. it breaks the official MDAnalysis tutorial Example: Saving dynamic per-atom properties in B-factor (which admittedly needs an overhaul...).
3. it used to to work... so it's a regression

[orbeckst]

Note: at the moment you have to do manually add the tempfactors topology attribute

u.add_TopologyAttr(MDAnalysis.core.topologyattrs.Tempfactors(np.zeros(len(u.atoms))))

u.atoms.tempfactors = 0.5
u.atoms.write("withbfacs.pdb")

[orbeckst]

We might not do much more here than properly document it. Opinions?

[richardjgowers]

@orbeckst I think it is possible to make this behaviour work, I (or someone else) just needs to get the hacking done under the hood. Short term documenting it is definitely a good thing though.

[weiHelloWorld]

@orbeckst Just curious why "b-factor" is renamed "tempfactor" in the new version?

[orbeckst]

tempFactor is the offical name in the ATOM record of the PDB format. We don't like camelCase so we use tempfactor. We also use the convention that "arrays" are plural words, hence tempfactors.

The new usage is incompatible with previous versions of MDAnalysis where you could do things like atoms.set_bfactors. But that was a specialized hack. The new system is universal and fully internally consistent so it was worth breaking the old behavior but of course we know that it is always a hassle for users – sorry!

EDIT: And I think we didn't mind renaming bfactors -> tempfactors because we were breaking everything anyway so we opted to go with the "proper" PDB name.

[weiHelloWorld]

I see, thank you!

[orbeckst]

This should work now with #1092 and is also documented under Universe.add_TopologyAttr.

[orbeckst]

@weiHelloWorld – just fyi, we added bfactors back... see the example that I linked above.

[weiHelloWorld]

@orbeckst Thank you!