Dear Partners,

we integrated an TUM in-house algorithm for structural clustering into the OpenTox framework. It is based on the graph mining tool gSpan' and much more efficient than related methods that are mostly based on MCS.
To get to know more about it please have a look at Stefan's slides from Rhodes (http://opentox.org/home/documents/presentations/rhodesworkshoppressept2010/at_download/file, slides 142, 148), at our ECML publication from this year (http://wwwkramer.in.tum.de/people/research/pubs/onlineGraphClustering) or our poster from the EuroQSAR2010 (http://wwwkramer.in.tum.de/people/people/girschic/resolveUid/1dbb03a8e183cc420975dac5d6e54585).
 

With the following two curl calls you can cluster data and apply a clustering model.

Cluster data / Learn a clustering model (structural overlap between molecules is set to be at least 80% in this example):
curl -X POST -d 'dataset_uri=http://apps.ideaconsult.net:8080/ambit2/dataset/3553' -d 'threshold=0.80' http://opentox.informatik.tu-muenchen.de:8080/OpenTox-dev/algorithm/StructuralClustering
---> http://opentox.informatik.tu-muenchen.de:8080/OpenTox-dev/model/TUMOpenToxModel_StructuralClustering_1

>From the RDF representation of the model (http://opentox.informatik.tu-muenchen.de:8080/OpenTox-dev/model/TUMOpenToxModel_StructuralClustering_1) the structural clusters can be read out, e.g.
<ot:dataset>
<ot:Dataset rdf:about="http://apps.ideaconsult.net:8080/ambit2/dataset/74011">
<ot:hasSource rdf:datatype="http://www.w3.org/2001/XMLSchema#string">Structural Clustering cluster 5</ot:hasSource>
</ot:Dataset>
</ot:dataset>



Use the learned clustering model for cluster assignment of test molecules:
curl -X POST -d 'dataset_uri=http://apps.ideaconsult.net:8080/ambit2/dataset/19629' -d 'threshold=0.50' http://opentox.informatik.tu-muenchen.de:8080/OpenTox-dev/model/TUMOpenToxModel_StructuralClustering_1
---> http://apps.ideaconsult.net:8080/ambit2/dataset/74195


If you have any questions please do not hesitate to contact us.

I wish you all a Merry Christmas and a Happy New Year!

Best regards,

Fabian
-- 
Neu: GMX De-Mail - Einfach wie E-Mail, sicher wie ein Brief!  
Jetzt De-Mail-Adresse reservieren: http://portal.gmx.net/de/go/demail