From: Martin Eklund <martin.eklund@fa...> - 2006-03-10 17:18:22
Sorry that I haven't replied until now.
The problem here is that the PDB-format is REALLY badly defined. In for
instance 1fdy the problem is that there are HETATMs in the middle of the
ATOM records. In 1tnl that the OXT-atom is not the last one in the file
(as it is supposed to be). However, to be honest, the PDB-parser in CDK
is witten in quite a bad way - it assumes quite a rigid file format. To
account for all the arbritariness that the PDB-format allows for, the
CDK PDB parser needs rather big extentions. I'm sorry, but I don't have
time to make these extentions now. The best thing would probably be to
write a general parser built by a parser generator, e.g. JavaCC or
ANTLR. If you are interested to have a go at extending the PDB-reader
yourself or indeed write a better parser with ANTLR, I can definitely
help in assisting you.
On Mon, 2006-03-06 at 15:56 +0100, sjoseph@... wrote:
> Hello Martin,
> I am student from Cologne university,who is using PDBReader in CDK and
> developing a scoring function.The PDBReader is throwing null pointer exception
> in function createbonds for some PDBIDS like
> etc.The number of files throwing exception is large that i have to debugg the
> error someway.
> The error java.lang.NullPointerException
> is in if(strand.getAtomCount() < atoms &&
> line in createBonds().
> The same kind of error occurs for files with OXT like 1tnl
> and without oxt like 1fdy.
> Sunu Joseph
Department of Pharmaceutical Biosciences
Uppsala University, Sweden