We provide access to some of the text mining capabilities of the OntoGene sytem through RESTful web services. Fundamentally you can send any document (or collection of documents) in BioC format and have it processed by the OntoGene pipeline.
BioC is a new standard for data exchange for text mining systems. It was developed at NLM/NCBI as part of an international collaboration (the OntoGene group is one of the members). It provides a simple format that can be used, among other things, to query a web service. BioC is defined by a simple DTD.
At the moment (September 4, 2014) the following services are available:
https://pub.cl.uzh.ch/projects/BioC/chem identifies chemical entities
https://pub.cl.uzh.ch/projects/BioC/chem+offset as above, with offsets (position of the term in the document)
https://pub.cl.uzh.ch/projects/BioC/disease identifies diseases
https://pub.cl.uzh.ch/projects/BioC/disease+offset as above, with offsets
https://pub.cl.uzh.ch/projects/BioC/gene identifies genes
https://pub.cl.uzh.ch/projects/BioC/gene+offset as above, with offsets
https://pub.cl.uzh.ch/projects/BioC/ctd delivers all the entities above (genes, chemicals, diseases)
https://pub.cl.uzh.ch/projects/BioC/ctd+offset as above, with offsets
https://pub.cl.uzh.ch/projects/BioC/ixn identifies "action terms" (which describe an interaction type)
https://pub.cl.uzh.ch/projects/BioC/ctd+rel delivers all the entitites (gene, chemical, diseases) and their interactions
https://pub.cl.uzh.ch/projects/BioC/ctd+offset+rel delivers relationships and entities, with offsets
Instructions on how to use the OntoGene RESTful service
In order to use the services, you need to submit a document (or collection) in BioC format (see example here) as a POST request to one of the services above.
There are different utilities that allow the submission of a document as a post request to a remote service. For example, you could use the curl utility and choose one of the following options:
curl SERVICE | curl -d "$(cat FILENAME)" -X POST
cat FILENAME | curl SERVICE -d @- -X POST
If your document is available from a web server, you can submit it for example as follows:
curl DOCURL | curl SERVICE -d @- -X POST
In the examples above, SERVICE is the URL of one of the OntoRest services (see above), DOCURL is the location of a remote document, FILENAME is the path of a local file. Beware that the local file, or the document retrieved at DOCURL, must be in BioC format.
You can either send a single BioC document to an OntoRest service (see example here), or a collection of documents (see example here).
For simple testing you can also use a browser plugin, such as "Poster" for Firefox. Once installed, you have the enter the url of the OntoRest service that you want to use, and attach the file that you want to analyze; then press the [POST] button.
In case of problems, please contact Fabio Rinaldi (fabio AT ontogene.org).
ODIN web interface
Additionally, you can test the OntoGene text mining capabilities through our web interface called ODIN (OntoGene document inspector). Currently two versions are publicly available:
ODIN CTD (customized for the Comparative Toxicogenomics Database)
ODIN RegulonDB (customized for the RegulonDB database)
The implementation of OntoRest is based on PyBioC, a python based implementation of the BioC standard developed at the University of Zurich. For technical details, contact Hernani Marques Madeira (h2m@access.uzh.ch).
Additional information available in the following publications:
Marques, Hernani; Rinaldi, Fabio (2013). PyBioC: a python implementation of the BioC core. In: Fourth BioCreative Challenge Evaluation Workshop, Bethesda, MD, US, 7 October 2013 - 9 October 2013, 2-4.
Fabio Rinaldi, Simon Clematide, Hernani Marques, Tilia Ellendorff, Martin Romacker, and Raul Rodriguez-Esteban. OntoGene web services for biomedical text mining. BMC Bioinformatics 2014, 15(Suppl 14):S6 doi:10.1186/1471-2105-15-S14-S6