Menu
▾
▴
Re: [Gusdev-gusdev] Relation between tables
|
From: <tw...@cs...> - 2004-03-12 14:48:30
|
Thomas, Could you send a pointer to the KEGG FTP URL you are accessing and the names (or type of) files that you are using? I would appreciate this to see how others are using KEGG. OMIM is orthogonal; I just described it and the parts of KEGG I addressed to explain the material on my web site. Terry On Fri, Mar 12, 2004 at 11:08:29AM -0300, Thomas Otto wrote: > Terry, > > first, I parse the information from the files of the ftp-server. I do > not use the xml stuff. > > True, there is a lot of stuff in kegg, but I want to represent it in > GUS, so... > > ... I think a link to OMIM makes sense, also to Motif/prosite. For me it > is important to have all the relations without redundend. > > Give me some time, I will see, which parts might help me, > > cheers, > Thomas > > > Terry Clark wrote: > > >Thomas, > > > >A few months ago I worked out a preliminary set of tables > >for *some* KEGG pathway data. > > > >(By the way, those reactions are not complete in some of > >the XML files.) I expect you know, there are two representations > >in the XML: one with proteins as nodes; the other with reaction > >products as nodes and proteins as edges. > >I focused on explicitly representing the latter > >(the <reaction> tag part of the XML) but have toyed with > >the idea to do a comprehensive KEGG representation in GUS unless > >someone else does it first :-) > > > >I also have a plugin to load the tables. All I am discussing has > >been implemented as a site-specific/local extension to GUS. > >I put the database files (not the plugin - we can talk > >about that later if you are interested) on the web page > > > > http://flora.uchicago.edu/gus/keggschemadraft/ > > > >One file you might check first at the URL is > > ** kegg-tables.sql *** > > > >I am not satisfied with some of the design, and there may > >be some uninspired things there - caveat emptor. I was > >waiting for KEGG to fix some problems in the reaction tags, > >and put together this operational db prototype in the meantime. > >(For example, check the missing - as of three weeks ago still - > >reaction tags in the XML for phenylalanine metabolism.) > > > >Anyway the files I point to above might be useful to > >start with rather than from scratch. > > > >I also hacked together some table(s) for EnzymeCommission > >numbers; there might be something in GUS for that, but at > >the time what I did seemed faster for my needs than otherwise. > >The OMIM parts there are not related to KEGG but part of > >a project I am working on. > > > > > >Terry > > > > > > > > > > > >On Thu, Mar 11, 2004 at 04:58:48PM -0300, Thomas Otto wrote: > > > > > >>I think I will all the relation in one table... sort of EnzymeRelations... > >> > >>Thomas > >>Thomas Otto wrote: > >> > >> > >> > >>>Okay, an example: > >>>To every enzyme are several reaction associated. i.e. Enzymea, has > >>>reaction1, ..., reaction n. > >>>(1) I could be to save the name of the reaction in the table > >>>EnzymeClassAttribute > >>>(2) would be, to save nothing of this in the EnzymeClassAttribute, but > >>>in a table (to create) EnzymeReactionRelation. > >>> > >>>This will mean a lot of new table, because just for the enzymes there > >>>are pathways, reaction, compound, substrate, product... > >>> > >>>But indeed, I also prefer #2, who is less redundend. > >>> > >>>Thanks, > >>>Thomas > >>> > >>>Angel Pizarro wrote: > >>> > >>> > >>> > >>>>In the absence of an example of data a definite schema, definitely #2 > >>>>as the data model of choice. You can always implement materialized > >>>>view to speed-up/ease queries. > >>>> > >>>>Cheers > >>>>Angel > >>>> > >>>>Thomas Otto wrote: > >>>> > >>>> > >>>> > >>>>>That's what I was thinking about. So I want to know, which > >>>>>representation of datas you prefer. (1 or 2, it is more a general > >>>>>question). > >>>>> > >>>>>Thomas > >>>>> > >>>>> > >>>>>Angel Pizarro wrote: > >>>>> > >>>>> > >>>>> > >>>>>>OK, I don't see where you are getting EnzymCompoundRelation & > >>>>>>CompoundClassAttribute from. Our current version of GUS does not > >>>>>>have these tables. The SRes.EnzymeClass table is also looking very > >>>>>>wacky to me. Propose a nice structure for this and we can put it in > >>>>>>the next release of GUS. > >>>>>> > >>>>>>Angel > >>>>>> > >>>>>>Thomas Otto wrote: > >>>>>> > >>>>>> > >>>>>> > >>>>>>>Hello, > >>>>>>> > >>>>>>>I am uploading some kegg data in the GUS system... > >>>>>>> > >>>>>>>So I looked which tables are existing, and which we are needing > >>>>>>>i.e. for compound, reactions. Now I am not sure how is the > >>>>>>>convention to relate the data. > >>>>>>> > >>>>>>>Example: > >>>>>>>The Enzymes are related to the compounds. > >>>>>>>(1) In EzymeClassAttribute I write the names of the compound > >>>>>>>related to each enzyme. In CompoundClassAttribute I write also all > >>>>>>>EC-numbers, related to it. > >>>>>>>(2) Other possibilty would be to utilize a thrid table > >>>>>>>EnzymeCompoundRelation, where I put the primkeys of the relations. > >>>>>>> > >>>>>>>(1) would be faster, easier to query, but redundend. > >>>>>>>(2) Cleaner > >>>>>>> > >>>>>>>What are you thinking I should use? > >>>>>>> > >>>>>>>Thanks, > >>>>>>>Thomas > >>>>>>> > >>>>>>> > >>>>>>> > >>>>>>> > >>>>>>>------------------------------------------------------- > >>>>>>>This SF.Net email is sponsored by: IBM Linux Tutorials > >>>>>>>Free Linux tutorial presented by Daniel Robbins, President and CEO of > >>>>>>>GenToo technologies. Learn everything from fundamentals to system > >>>>>>>administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click > >>>>>>>_______________________________________________ > >>>>>>>Gusdev-gusdev mailing list > >>>>>>>Gus...@li... > >>>>>>>https://lists.sourceforge.net/lists/listinfo/gusdev-gusdev > >>>>>>> > >>>>>>> > >>>>>> > >>>>>> > >>>>>> > >>>>>> > >>>>>> > >>>>>> > >>>> > >>>>------------------------------------------------------- > >>>>This SF.Net email is sponsored by: IBM Linux Tutorials > >>>>Free Linux tutorial presented by Daniel Robbins, President and CEO of > >>>>GenToo technologies. Learn everything from fundamentals to system > >>>>administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click > >>>>_______________________________________________ > >>>>Gusdev-gusdev mailing list > >>>>Gus...@li... > >>>>https://lists.sourceforge.net/lists/listinfo/gusdev-gusdev > >>>> > >>>> > >>>> > >>> > >>> > >>>------------------------------------------------------- > >>>This SF.Net email is sponsored by: IBM Linux Tutorials > >>>Free Linux tutorial presented by Daniel Robbins, President and CEO of > >>>GenToo technologies. Learn everything from fundamentals to system > >>>administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click > >>>_______________________________________________ > >>>Gusdev-gusdev mailing list > >>>Gus...@li... > >>>https://lists.sourceforge.net/lists/listinfo/gusdev-gusdev > >>> > >>> > >>> > >> > >> > >>------------------------------------------------------- > >>This SF.Net email is sponsored by: IBM Linux Tutorials > >>Free Linux tutorial presented by Daniel Robbins, President and CEO of > >>GenToo technologies. Learn everything from fundamentals to system > >>administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click > >>_______________________________________________ > >>Gusdev-gusdev mailing list > >>Gus...@li... > >>https://lists.sourceforge.net/lists/listinfo/gusdev-gusdev > >> > >> > > > > > > > |
