[OTDev] running Algorithm directly on a SD file?
Nina Jeliazkova nina at acad.bgFri Apr 2 10:17:23 CEST 2010
- Previous message: [OTDev] running Algorithm directly on a SD file?
- Next message: [OTDev] running Algorithm directly on a SD file?
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Hello Egon, Egon Willighagen wrote: > On Wed, Mar 31, 2010 at 12:27 PM, Nina Jeliazkova <nina at acad.bg> wrote: > >>> is it possible to run an OpenTox /algorithm directly on an SD or mol >>> file? Or should that always be uploaded as data set first? >>> >>> >> The dataset should have an URI, so it should be uploaded, or the dataset >> constructed from existing compounds/features at a certain dataset service. >> > > Must the URI point to an OpenTox API service? Or would a URI to just a > SD file work too? > Like http://example.org/someFile.sdf ? > For some algorithms it might do, but IMHO most of the services support only the mandatory RDF format so far. One note for processing working directly with files rather than URI - the latest was agreed within OpenTox, for 1) its best fit with the REST architecture - all resources (regardless of being data or models) are exposed by URIs 2) these dataset URIs are used to link algorithm, methods and features within RDF representation. If we drop the requirements of having URIs, it might be crucial for the consistency of all OpenTox framework. As an example ,see curl -H "Accept:application/rdf+xml" http://opentox.ntua.gr:3000/model/3 (or just http://opentox.ntua.gr:3000/model/3 in a browser ) <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:ot="http://www.opentox.org/api/1.1#" xmlns:ota="http://www.opentox.org/algorithmTypes.owl/#" xmlns:owl="http://www.w3.org/2002/07/owl#" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:xsd="http://www.w3.org/2001/XMLSchema#" xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#" > .... <ot:trainingDataset rdf:resource="http://ambit.uni-plovdiv.bg:8080/ambit2/dataset/269"/> .... Another example is the descriptor calculation. We are storing the URI of the descriptor calculation algorithm in the feature RDF, in order to be able to reuse it for new compounds. In the RDF representation of a dataset (below) there are links to the algorithm, used to generate descriptors , or models, used to predict properties. Unfortunately, all this "linked data" will break down, if submitting files, not URIs is introduced. curl -H "Accept:text/n3" http://ambit.uni-plovdiv.bg:8080/ambit2/dataset/607 or http://ambit.uni-plovdiv.bg:8080/ambit2/dataset/607 @prefix ot: <http://www.opentox.org/api/1.1#> . @prefix ad: <http://ambit.uni-plovdiv.bg:8080/ambit2/dataset/> . @prefix af: <http://ambit.uni-plovdiv.bg:8080/ambit2/feature/> . ad:607 a ot:Dataset ; ot:dataEntry [ a ot:DataEntry ; ot:compound <http://ambit.uni-plovdiv.bg:8080/ambit2/compound/101215/conformer/269405> ; ot:values [ a ot:FeatureValue ; ot:feature af:103141 ; ot:value "0.633"^^xsd:double ] ; ot:values [ a ot:FeatureValue ; ot:feature af:106946 ; ot:value "2.695"^^xsd:double ] ] . ot:hasSource a owl:ObjectProperty . af:103141 a ot:Feature , ot:NumericFeature ; dc:creator "http://ambit.uni-plovdiv.bg:8080/ambit2" ; dc:identifier "http://ambit.uni-plovdiv.bg:8080/ambit2/feature/103141"^^xsd:anyURI ; dc:title "TUM_CDK_XLogP" ; ot:hasSource <http://opentox.informatik.tu-muenchen.de:8080/OpenTox-dev/algorithm/CDKPhysChem/XLogPDescriptor> ; ot:units "" ; = ot:TUM_CDK_XLogP . af:106946 a ot:Feature , ot:NumericFeature ; dc:creator "http://ambit.uni-plovdiv.bg:8080/ambit2" ; dc:identifier "http://ambit.uni-plovdiv.bg:8080/ambit2/feature/106946"^^xsd:anyURI ; dc:title "Empty feature for prediction." ; ot:hasSource <http://opentox.informatik.tu-muenchen.de:8080/OpenTox-dev/model/TUMOpenToxModel_kNN_6> ; ot:units "" ; = ot: . Looking to your comments, Nina > Egon > >
- Previous message: [OTDev] running Algorithm directly on a SD file?
- Next message: [OTDev] running Algorithm directly on a SD file?
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Development mailing list