-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi Bernhard,

I am not sure this represents your problem: when I only select the
ATOM cards and of those only columns 1-56

grep "^ATOM" 9INS.pdb | cut -c 1-56 > 9INS_coords.pdb

the molprobity server has no problems analysing this file including
the addition of hydrogen atoms.

The PDB-file has lines such as
ATOM    433  N   LYS B  29      15.668  49.766  24.933
ATOM    434  CA  LYS B  29      14.854  50.940  25.268
ATOM    435  C   LYS B  29      13.545  50.606  25.960
ATOM    436  O   LYS B  29      12.781  51.552  26.227

i.e. no B-values, no occupancies, no CRYST1 etc.

Again, I am not sure this mutilation corresponds to the problem case
you describe.

Best,
Tim

On 05/21/2014 01:43 PM, Bernhard Rupp wrote:
> Dear All,
> 
> 
> 
> I experienced an odd observation with some philosophical
> consequences (here we go again)
> 
> 
> 
> When I run a model obtained from a perhaps popular but until
> further examination unnamed server,
> 
> and validate through coot/validate, the model kills probe/reduce
> which normally works great.
> 
> 
> 
> Following observations:
> 
> Normally, 3 windows appear: Generic objects, Flips, Molprobity
> Probe clash gaps; and the H-added model and the little dots and
> spikes are displayed.
> 
> For that model I get only the Flips window with 'no adjustment',
> and an empty Molprobity Probe Clash Gaps Window.
> 
> 
> 
> The model has only a CRYST1 record,
> 
> does NOT contain a B factor or occupancy column
> 
> it HAS a TER record.
> 
> 
> 
> OK, I probably can paste/awk those missing columns in there with
> fake values.
> 
> 
> 
> But now the consequences:
> 
> 
> 
> Also the PDB adit server rejects that model, although I cannot see
> why one should not be able to do at least a geometry
> 
> validation with coordinates alone.
> 
> 
> 
> Particularly the sucker models returned from structural
> bioinformatics servers (which understandably and in rare humbleness
> 
> 
> do not contain occs or Bs), need exactly this validation the most.
> The model in question has overlapping backbone sections and rmszs
> >70.
> 
> 
> 
> What means are left for a user to validate such stuff? If the PDB
> wants to be a true structure resource for users, it should allow
> such
> 
> minimum validation of any coordinate model, and not reject
> everything that does not fit stringent deposition criteria. I do
> not need
> 
> CRYST or TER records and not even B's or Occs  if I want to know if
> a model is really bad. I think that these weird models
> 
> that are spit out by certain bioinformatics servers are a lot more
> hazardous to an innocent user than a poor experimental model.
> 
> 
> 
> Cheers, BR
> 
> 
> 
> ----------------------------------------- Trigger warning 
> ---------------------------------------------------
> 
> This message may or may not contain references to issues of
> privilege and oppression
> 
> including but not limited to enantiophobia, point classism,
> heteroatomism, transbondism,
> 
> cisbonding, sizeism, curvature, references to color, alcohol,
> blood, small insects, disgusting cells, and
> 
> any combination thereof that may cause symptoms ranging from
> discomfort and anxiety to violent
> 
> physical response in sensitive individuals and must therefore be
> labeled as potentially
> 
> hazardous to your comfort and well-being.
> 
> 
> 
> 

- -- 
- --
Dr Tim Gruene
Institut fuer anorganische Chemie
Tammannstr. 4
D-37077 Goettingen

GPG Key ID = A46BEE1A

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)
Comment: Using GnuPG with Icedove - http://www.enigmail.net/

iD8DBQFTfKH6UxlJ7aRr7hoRAq56AKD03HICPNfcDlTGVyQvoQuasGUtHwCfdRLa
ZbOuyXr5Yylh6OyTLbz8ftE=
=SdJM
-----END PGP SIGNATURE-----

Reply via email to