1. BioNLP-SADI: BioNLP Semantic Web Services based on SADI FrameworkBioNLP-SADI: BioNLP Semantic Web Services based on SADI Framework
Ahmad C. Bukhari, Artjom Klein, and Christopher J.O. Baker*Ahmad C. Bukhari, Artjom Klein, and Christopher J.O. Baker*
Department of Computer Science and Applied Statistics, University of New Brunswick, Saint John, NB CanadaDepartment of Computer Science and Applied Statistics, University of New Brunswick, Saint John, NB Canada
IntroductionIntroduction
Over the last decade biomedical naturalOver the last decade biomedical natural
language processing (BioNLP) has been
validated as an solution to address the text-
mining and information extraction needs ofmining and information extraction needs of
life scientists. Recently the number of NLP
and BioNLP tools published as Web Servicesand BioNLP tools published as Web Services
(WS) has been growing steadily. BioNLP WSs
available to the BioNLP community include:available to the BioNLP community include:
Whatizit, e-LICO WSs, and text-mining WSs in
BioCatalogue. Web Services do not require
installation, are platform independent, and BioNLP SADI Services in Context
installation, are platform independent, and
provide access to programs that can not be
installed on regular computers due their complexity and heaviness. In many use-cases,installed on regular computers due their complexity and heaviness. In many use-cases,
integration of several text-mining WSs is required and the output results must be
consolidated. Most of BioNLP tools produce XML or TAB-output (with different
schemas), where the integration of WSs and consolidation of results requiresschemas), where the integration of WSs and consolidation of results requires
programming.
We leverage Semantic Web Services that combine Web Service technology with theWe leverage Semantic Web Services that combine Web Service technology with the
intelligent processing of ontology-based metadata to achieve application and data
integration scenarios. We propose a programming and installation free scientific textintegration scenarios. We propose a programming and installation free scientific text
processing system to annotate and to extract biologically valuable information from
textual data based on SADI framework. The basic purpose of this solution is to create a
web based platform for the bio text-miner and more specifically the bioinformatics
Proposed Methods
web based platform for the bio text-miner and more specifically the bioinformatics
application and database developer with minimal extra effort.
Proposed Methods
1. Service output represented in RDF format.1. Service output represented in RDF format.
2. OWL ontologies for modeling to ensure interoperability of the web services.
3. Semantic triples stores to store, query, and manipulate text mining results.3. Semantic triples stores to store, query, and manipulate text mining results.
4. SPARQL query language for ad-hoc semantic querying on results and
implementation of benchmarking evaluation metrics for comparative evaluation
of web services.of web services.
5. SADI SPARQL Clients for ad-hoc query and consolidation of text-mining results
with data from biological databases.with data from biological databases.
6. Third-party tools and APIs for easy access (Taverna, Web Interface, SADI Java,
Python, and Perl APIs, Annotation Toolkits with graphical interface).Python, and Perl APIs, Annotation Toolkits with graphical interface).
Benefits
Anatomy of BioNLP-SADI
Benefits
1. Ad-hoc consolidation of text mining results1. Ad-hoc consolidation of text mining results
2. Construct text mining pipelines for users without programming skills
3. Easy comparative evaluation of BioNLP tools3. Easy comparative evaluation of BioNLP tools
4. Compatibility with semantic data web services
SADI: BioNLP Semantic Web Services based on SADI FrameworkSADI: BioNLP Semantic Web Services based on SADI Framework
Ahmad C. Bukhari, Artjom Klein, and Christopher J.O. Baker*Ahmad C. Bukhari, Artjom Klein, and Christopher J.O. Baker*
Department of Computer Science and Applied Statistics, University of New Brunswick, Saint John, NB CanadaDepartment of Computer Science and Applied Statistics, University of New Brunswick, Saint John, NB Canada
Implemented Services and Target Queries
SADI
SADI Services Target Queries
Implemented Services and Target Queries
SADI Services Target Queries
Drug-Extraction
Drug-Drug-Interaction
Extract all drug names in a text and list interactions
(if they exist) between each of the drug pairs.Drug-Drug-Interaction (if they exist) between each of the drug pairs.
Food-Extraction
Drug-Food-Interaction
List the foods which have negative interaction
with the drug Cytarabine.Drug-Food-Interaction with the drug Cytarabine.
Mutation-Finder
Drug-Extraction
Find sentences where mutation and drug name occur
Drug-Extraction
Sentence-Splitter
Find sentences where mutation and drug name occur
in the same sentence.
Mutation-Finder Find abstracts where the same mutation mention isMutation-Finder
Sentence-Splitter
Find abstracts where the same mutation mention is
in two adjacent sentences.
Service Interoperability and
Ad-hoc Output ConsolidationAd-hoc Output Consolidation
Drug-Extraction and Drug-Drug-Interaction SADI services Consolidated Output
Summary
To address the ongoing challenges of poor interoperability and integration among
biological NLP services, we proposed a generalized architecture for a suite of SADI
services permitting the extraction of biological names entities from scientificservices permitting the extraction of biological names entities from scientific
documents and the formatting of extracted information in RDF for facile interoperable
reuse of the data. In future work, we will integrate more BioNLP tools and develop usereuse of the data. In future work, we will integrate more BioNLP tools and develop use
cases that illustrate the benefit of interoperable text mining and SADI data mining web
services.
References
services.
References
1. Wilkinson, M.D., Vandervalk, B., McCarthy, L.: The Semantic Automated Discovery and Integration
(SADI) Web service Design-Pattern, API and Reference Implementation. Journal of Biomedical
Semantics 2, 8 (2011)Semantics 2, 8 (2011)
2. Dietrich, R.S., Miguel, A., Sylvain, G., Harald, K.,Antonio, J.: Text processing through web services:
Calling whatizit. Bioinformatics, 24(2):296–298, November 2007.Calling whatizit. Bioinformatics, 24(2):296–298, November 2007.
3. Khalid, B., Franck, Tanoh, et al. Biocatalogue: A Curated Web Service Registry for the Life Science
Community. 3rd Intl Conference of BioCurators (2009).
4. e-LICO text-mining web services: http://www.e-lico.eu/text-mining-ws.html4. e-LICO text-mining web services: http://www.e-lico.eu/text-mining-ws.html
5. Hull, K., Wolstencroft, R., Stevens, C., Goble, M., Pocock, P.L.: Taverna: a tool for building and
running workflows of services. Nucleic Acids Research, vol. 34, iss. Web Server issue, pp. 729-732,
2006.2006.
6. Klein, A., Riazanov, A., Hindle, M.M., and Baker, C.J.O. Benchmarking infrastructure for mutation
text mining, Proc. AIMM2012, 2012
Project Page: https://code.google.com/p/bionlp-sadi/