[Cod-bugs] _atom_site_disorder_group

[Andrius Merkys]

Hi Marcin,

Thanks a lot for the interesting observation. I have opened an issue in
the COD's issue tracker [1] to deal with non-numeric values of
'_atom_site_disorder_group'.

I have looked into a bunch of such CIF files, and fixed a couple of
them. It seems that sometimes the problem is incorrect use of CIF data
types, and in some cases the issue is due to typographic mistakes.

We'll consider implementing checks in our software. Fixes mostly will
have to be done manually.

Best wishes,
Andrius

[1] https://projects.ibt.lt/repositories/issues/537

On 2020-03-02 11:57, Marcin Wojdyr wrote:
> last week I exchanged a few emails with Daniel Kratzert about handling
> _atom_site_disorder_group. Normally, the value is an integer, but it
> has type 'char' in the dictionary.
>
> We checked non-numeric values in COD. The full list is attached.
> Daniel thought that maybe SHELXL used letters (A, B) for
> _atom_site_disorder_group for a short time in 1993, but G. Sheldrick
> didn't remember it.
> Anyway, half of the cases with non-integer disorder group seems to be a mistake.
> Often a character is added to the dot (.. or \. or ?. or ...).
> In 2007289 it's repeated type_symbol. In 2005194 it's occupancy?
>
> For our use, we decided to just ignore all non-integer disorder
> groups, as they are rare and unreliable. I don't know what COD should
> do with it, but perhaps this list will be useful to you.

-- 
Andrius Merkys
Vilnius University Institute of Biotechnology, Saulėtekio al. 7, room V325
LT-10257 Vilnius, Lithuania



-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.