From owner-metabolic-reg@net.bio.net Wed Sep 22 01:03:00 1999 Path: biosci!newsfeed.stanford.edu!skynet.be!newsfeed.gamma.ru!Gamma.RU!demos!news.stack.serpukhov.su!not-for-mail From: "Alex Sel'kov" Newsgroups: bionet.metabolic-reg Subject: Re: Universal format for storing pathways and models Date: Wed, 22 Sep 1999 04:56:58 +0400 Organization: Not organized Lines: 119 Message-ID: <937958123.141954@sirius.shome.eu.org> References: <40EF164E4C27D31192760000F847DF32519954@ukz206.ggr.co.uk> <937692632.667327@sirius.shome.eu.org> <7s216b$ug5$1@nnrp01.ops.uunet.co.za> <937750899.40114@sirius.shome.eu.org> <937852968.19861.0.nnrp-07.c2de3927@news.demon.co.uk> <937860976.418122@sirius.shome.eu.org> <937899606.7504.0.nnrp-08.c2de3927@news.demon.co.uk> NNTP-Posting-Host: sirius.stack.net X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 5.00.2314.1300 X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2314.1300 Cache-Post-Path: sirius.shome.eu.org!unknown@lamb.empproject.com X-Cache: nntpcache 2.3.3 (see http://www.nntpcache.org/) > You should put the EMP spec on the net. Can you give a very small example, > eg an example to specify the following model: > > Note: X0 and X1 are boundary metabolties. > > X0 -> S1; k1*X0; > 2 S1 -> X1; k2*S1; > > k1 = 1.2; k2 = 4.5; > S1 = 0; X0 = 1.0; X1 = 0.0; > I think that looking at this maybe slightly better way to describe our situation: 1. EMP record like http://wit.mcs.anl.gov/EMP/emp_by_id.cgi?&id=SEL94023-01 2. Pathway http://wit.mcs.anl.gov/WIT2/CGI/pw.cgi?pathway=MLTPIBG6P.CAT&reforg=&user=&request=show_diagram 3. Related mechanism http://wit.mcs.anl.gov/ERM/diagram.cgi?id=MPPBB We only try to be useful datasource. So we need formal and useful way to represent original publications and different ways to export our data and nets. I can't say that our current formats are satisfactory and we definitely need users feedback and good ideas. > Perhaps we should think about a net wide effort? > Yes, I think so. > Not that I know of. What are Boehringer pathways (apart fom the wall > charts)? Actually the same wall charts. But thanks to recent tecnology improvments, you can use it as nice desktop wallpapers too. > We have thought of other software packages. One reason for opting for a > binary standard was that it would be easy for anyone to write a parser to > read and export pmb files. So we thought that DBSolve for example could > import and export pmb files, it would probably only take a day or two to > write such a utility. If the standard were structured ascii it would take > much longer for someone to write an importer, possibly weeks or more (syntax > checking etc). The pml standard we are developing will be able to export pmb > files. Note that the pmb file is fully extensible and portable, it has been > designed so that any changes made to the standard will not make old pmb > files redundant. It's OK. I perfectly understand your points. But I need good format for storing models in an universal manner. Difference like between e-mail text and web :) By example, I need fully standard model names, constant names, all beings which have names must keep their names in my nomenclature files :) > There is some metabolic visualisation software which I've written called > Talis, see my web site for details. I hope that Talis will be availiable soon for download and preview. > I'm not such interested in database (I > work with them all day anyway commerically) work, I'm much more keen on > metabolic dynamics and engineering and understanding how such complex > systems work. Trawling through database data and doing sql queries is not > that interesting to me. We're in the process of defining a human readable > spec for defining and manipulating metabolic pathways, the so-called pml > standard. It's very interesting for us. Espetially definition of pathway, principles of composition/decomposition/outlining, abstract net objects, principles of pathway naming e.t.c. You can find our current pathway editor prototype written by Gene Selkov at http://www.xnet.com/~selkovjr/ElectricArc/ if you interested. > What I think is more important from my point of view is enabling > 'ordinary' biochemists and molecular biologists to write and model systems > without having to learn an unfamiliar syntax such as XML etc. Instead I > think one should opt for a simple basic like language which everyone will be > familiar with. And from my point of view, biochemists and biologists must have searchable knowledge base where he/she can find other models and data :) > > I can say only one thing - this solution is perfectly fits if you need > exchange with > > a very few model by occassion. But when we speak about thousands of models > > with many various tools working simuntaniously on the same data - you > can't > > base your work on formats like pbm. Definitely not. > > > But the spec can exchange large models too, there is no limit to the size of > model one can specifiy. > I don't ever think about size and complexity of one model. What about 10000 different models and data sets for this models in one searchable database? Or you think that such beast will never exist? > > In any case I can't say that pbm is "bad" or "good". Simply such format > > unfotutantely can't solve our (by example) problems. > > What precisely is the aim of the work you do? We try to act as good informational source. And search and keep any facts and relations between objects. We have more 26,000 records and ~3000 maps. And how do you think, by example, how many possible ways to express enzyme activity exists, thanks to authors? You tell me about good basket for fruits and I immediately start to think about apples, oranges, eggs and old boots which will be kept in this container... Regards, as .