[Gusdev-gusdev] KEEG to GUS

[Thomas O. <ot...@fi...>]

Hi,
                                                                                                     
 
I think it is time to discuss a bit about the stuff in did so far 
concerning Kegg to Gus.
                                                                                                     
 
The principle idea was to put the datas in a relations database. The 
visualization of the pathway will not be changed.
                                                                                                     
 
So we did the schema for the oracle, under the namespace CLASS. Some 
tables already exist, like pathway, OMIM, enzymeclass... (but we kept it 
in a package). Still Trish and I are talking about the tail of the 
pathway-table (interaction and so). Shouldn't be too difficult to parse 
the information from the tables for the compounds, reactions and enzymes 
to the interaction tables.
                                                                                                     
 
Some might say this approve is a little bit static, but at least it a 
base of informations.
                                                                                                     
 
We also wrote a parser (a bit beta stage, but coming soon) to upload the 
data in the tables. The data itself can be downloaded at 
ftp://ftp.genome.ad.jp/pub/kegg. Further we are working on a 
classification for protein and functions.
                                                                                                     
 
To use the pathways locally, we wrote some scripts to surf through the 
pathway, the enzyme, compounds and so one, looks like under kegg, just 
in perl (do we need this in C?), so a bit slower and not complete so far 
(i.e. structure information). If there is interest, I will install it on 
a webserver. (If you want to be a betatester :-))
                                                                                                     
 
Maybe it would be useful to talk first about the tableschema before to 
get from alpha stage to the beta with the programms.
We hope to get some new ideas.
                                                                                                     
 
Cheers,
Thomas
                                                                                                     
 
PS: All stuff is in beta stage (I couldn't use the dumpSchema.pl :-(), 
so please tell me the bugs, thanks.