Dear Roberto,
Roberto Steiner wrote:
> Dear all,
>
> a problem possibly at the coot/mmdb interface...
Indeed.
> If one uploads a pdb file (from phenix.refine in the example below)
> that contains Hs into Coot and then writes it out (with or without any
> modification done on it) Coot shifts the HG2n of THR on the right by
> one column space. Because column 17 is kept empty the result is three
> identical HG2 THR protons.
Technical answer:
The current PDB parser in Coot is from mmdb and is for PDB format
version 2.3. Phenix.refine and other modern programs use PDB format
version 3.0 [1] - released over a year ago. PDB version 3.0
"Remediated" does not wrap the hydrogen (or other) names and so the
hydrogen name "unmangler" - which is what is tripping you up - need not
be executed. I'd like to make Coot compatible with the current PDB
standard. If CCP4 were to release a version of mmdb compatible with
Coot, I could do that right away [2]. I'm hoping that they will do so
today^H^H^H^H^H soon.
> The only solution right now (that I know of) is to remove all Hs and
> generate them again (molprobity for example) prior to refinement.
Bleugh. Non-optimal (no matter how fine Molprobity is).
Paul.
[1] or version 3.1?
[2] I imagine.
|