Bad geometry leaves histidine HE2 unbonded on PDB load

[Tristan Croll]

This is a fairly artificial situation, but nonetheless...

As a test of how well ISOLDE can recover from bad geometry, I created a deliberately badly distorted model (after adding hydrogens with AddH):

import numpy
m = session.models.list()[0]
m.atoms.coords += numpy.random.rand(*m.atoms.coords.shape)-0.5

If I save the resulting model as a PDB file and reload, the histidine HE2 atoms are no longer attached - I guess because the doubly-protonated His isn't in the dictionary of standard residues?

[Eric Pettersen]

I wouldn't normally bother with this kind of problem since bulletproofing ChimeraX against non-physical system is basically last on the list of priorities, but the fix was easy: have HIS residues default to using the HIP template instead of the HID template.