View
215
Download
0
Category
Tags:
Preview:
Citation preview
The Role of Ontologies in Improved Scholarly
Communication
Philip E. BourneUniversity of California San Diego
pbourne@ucsd.eduhttp://www.sdsc.edu/pb
My Perspective …• Ontology Developer (years ago – mmCIF -
Bioinformatics 2002 18: 1280-128)• Database Developer – RCSB PDB• Supporter of open access (provided there is a
business model) - editor in chief of PLoS Computational Biology
• Co-founder - SciVee Inc. • I am becoming increasingly interested in scholarly
communication• I use ontologies to support this work
Objective Today
• Describe how we are using ontologies to try and improve scholarly communication
• Motivate you towards thinking about ontologies that should be developed
• Learn from you where we might spend our efforts
First Consider What Motivates Us to Improve Scholarly
Communication
We Cannot Possibly Read a Fraction of the Papers We Should
Drivers of Change Renear & Palmer 2009 Science 325:828-832
Hence We Are Scanning More Reading Less
Renear & Palmer 2009 Science 325:828-832Drivers of Change
The Truth About the Scientific eLaboratory
• I have ?? mail folders!
• The intellectual memory of my laboratory is in those folders
• This is an unhealthy hub and spoke mentality
Drivers of Change
The Truth About the Scientific eLaboratory
• I generate way more negative that positive data, but where is it?
• Content management is a mess– Slides, posters…..– Data, lab notebooks ….– Collaborations, Journal clubs …
• Software is open but where is it?• Farewell is for the data too
Drivers of Change
Computational Biology Resources Lack Persistence and Usability. PLoS Comp. Biol. 4(7): e1000136
Data and the Publication Are Disjoint
• PubMed contains 18,792,257 entries
• ~100,000 papers indexed per month
• In Feb 2009:– 67,406,898 interactive
searches were done– 92,216,786 entries were
viewed
• 1078 databases reported in NAR 2008
• MetaBase http://biodatabase.org reports 2,651 entries edited 12,587 times
Biosciences Data as of April 14, 2009Drivers of Change
Publishing Limitations
• A paper is an artifact of a previous era• It is not the logical end product of eScience,
hence:– Work is omitted– Article vs supplement is a mess– Visualization may be limited– Interaction and enquiry are non-existent– Rich media can help, but are rarely used
Drivers of Change
We Need to do Better & The Game is Afoot
It is being driven from the top down and the bottom up
Ontologies & Semantic Tagging
BioLit Data Extraction/StorageDatabase IDsOntology termsText excerptsOther… BioLit
MySQLdatabase
XML
XML,Meta-data
<w
eb
se
rvic
es>
we
b
ext
ern
ald
ata
bas
es
Semantic Tagging
Tagging of PubMed Central
• Ontologies read from OBO Files• Words converted to tree structures• Matched to every non-trivial word in the
paper• Matches tagged• A long paper can be matched to GO in less
than 30 seconds
Semantic Tagging http://biolit.ucsd.edu
Semantic Tagging http://biolit.ucsd.edu
ICTP Trieste, December 10, 200716
http://biolit.ucsd.eduSemantic Tagging
Provision of Webservices to this tagging may be the most valuable contribution..
Semantic Tagging
www.rcsb.org/pdb/explore/literature.do?structureId=1TIMDatabase & Literature Integration
Context
BMC Bioinformatics 2010 11:220Semantic Tagging
Semantic Tagging of Database Content
http://www.pdb.orgPLoS Comp. Biol. 6(2) e1000673Semantic Tagging
Automatic Knowledge Discovery for Those with No Time to Read
Immunology Literature
Cardiac DiseaseLiterature
Shared FunctionSemantic Tagging
This is Literature Post-processingBetter to Get the Authors Involved
• Authors are the absolute experts on the content
• More effective distribution of labor
• Add metadata before the article enters the publishing process
BMC Bioinformatics 2010 11:103Semantic Tagging
Word 2007 Add-in for Authors
• Allows authors to add metadata as they write, before they submit the manuscript
• Authors are assisted by automated term recognition– OBO ontologies– Database IDs
• Metadata are embedded directly into the manuscript document via XML tags, OOXML format– Open– Machine-readable
• Open source, Microsoft Public License
http://www.codeplex.com/ucsdbiolitDrivers of Change
Word 2007 Add-in Example of What it Looks Like - Ontologies
• Inline Recognition, Highlighting, and Mark-up of Informative Terms– A recognized term will have a dotted, purple underline– Hovering generates a Smart Tag above the term
• add mark-up for this term• ignore this term• view the term in the ontology browser• If a recognized term appears in more than one ontology, all instances
of that term will be listed– Hovering over a marked-up term
• option to apply mark-up to all recognized instances of term• stop recognizing a term
– Pass ontology terms back to provider
Semantic Tagging BMC Bioinformatics 2010 11:103
• Built-in Knowledge of Ontologies and Databases– Add-in provides a list of biomedical ontologies to
download– and a list of databases for ID recognition
(GenBank/RefSeq, UniProt, Protein Data Bank)– A user may also supply a URL to download other
ontologies
• Ontology Browser– allows a user to select an ontology and then navigate
through it to view terms and their relationships
BMC Bioinformatics 2010 11:103
Custom Metadata• Ontologies do not contain all usages of a concept• Add-in allows user to assign custom metadata
• Human Disease Ontology term: Leukemia, T-Cell, HTLV-II-Associated
• Synonym: Atypical hairy cell leukemia (disorder) • Actual use in literature:
– hairy cell leukemia– hairy-cell leukemia– hairy T cell leukemia– T cell hairy leukemia
BMC Bioinformatics 2010 11:103
Synonym mapping, disambiguation
• Inclusion of an additional set of synonyms for a term that reflect its use in natural language– Automated finding of synonyms in extant
literature– Gather synonyms from term-mapping databases
• Incorporate a more sophisticated term recognition approach into the add-in
BMC Bioinformatics 2010 11:103
Challenges
• Author use– Familiarity with ontologies, terms– Agreement between co-authors
• End-use of semantically enriched manuscript
• Need to combine with NLM XML standard
Semantic Tagging BMC Bioinformatics 2010 11:103
Challenges: Author Use
IF one or more publishers fast tracked a paper that had semantic
markup I would argue it would catch on in no time
Semantic Tagging BMC Bioinformatics 2010 11:103
Where we Need {Better} Ontologies
1. To Support Mashups Between Different Types of Scholarly Output
Post-publication of Video and Paperwww.scivee.tv
Drivers of Change
Pubcast – Video Integrated with the Full Text of the Paper
Pubcasts - A Unique Technology
Don’t understand what you are reading? Click and have the author pop-up and explain it!
See the scientists and the experiments behind the research papers and textbooks
Pubcasts - A Blend of Video, text, tables, figures, PowerPoints, comments, ratings…ALL SYNCHRONIZED FOR RAPID LEARNING
Mashups – www.scivee.tv
Where we Need {Better} Ontologies
2. To Support Tagging of all Aspects of the Scholarly Product
Consider Today’s Academic Workflow
Research[Grants]
JournalArticle
ConferencePaper
PosterSession
Feds
Societies
Publishers
Reviews
BlogsCommunity Service/Data
Curation
What Should be Done?
Consider Tomorrow’s Academic Workflow
Research[Grants]
JournalArticle
ConferencePaper
PosterSession
Feds
Societies
Publishers
Reviews
BlogsCommunity Service/Data
CurationIdeas, Data, Hypotheses
What Should be Done?
Maybe The Line is Somewhere Else?
Scientist
Idea
Experiment
Data
Conclusions
Publish
Laboratory
Publisher
Maybe The Line is Somewhere Else?
Scientist
Idea
Experiment
Data
Conclusions
PublishWhat Should We Do?
Laboratory
Publisher
Institution
Lab Notebook
Crowd Sourcing the Electronic Printing Press(aka Workshop: Beyond the PDF)
• Proposal to the US National Science Foundation:
• Aims:– Define user requirements– Establish a specification document– Open source the development effort– Have a commitment from a publisher to publish a
research object using the system– Act as an exemplar for what can be done
Question: What if Everyone Had An Electronic Printing Press?
• Peer review might change?• Bibliometrics might change?• Business models will likely change?• What happens to the database/literature divide?• Societies might do more self publishing?• We might have improved the dissemination of
science, but will we have improved the comprehension?
General References
• What Do I Want from the Publisher of the Future PLoS Comp Biol http://www.sdsc.edu/pb
• Fourth Paradigm: Data Intensive Scientific Discovery http://research.microsoft.com/enus/collaboration/fourthparadigm/
References to Exemplars
• Semantic Biochemical Journal - 2010: Using Utopia
• Article of the Future, Cell, 2009:• Prospect, Royal Society of Chemistry, 2009:• Adventures in Semantic Publishing, Oxford U, 2009:
• The Structured Digital Abstract, Seringhaus/Gerstein, 2008• CWA Nanopublications – 2010
Acknowledgements• BioLit Team
– Lynn Fink– Parker Williams– Marco Martinez– Rahul Chandran– Greg Quinn
• Microsoft Scholarly Communications– Pablo Fernicola– Lee Dirks– Savas Parastitidas– Alex Wade– Tony Hey
• wwPDB team
• SciVee Team– Apryl Bailey– Tim Beck– Leo Chalupa– Lynn Fink– Marc Friedman (CEO)– Ken Liu– Alex Ramos– Willy Suwanto
http://www.scivee.tv
http://biolit.ucsd.eduhttp//www.pdb.orghttp://www.codeplex.com/ucsdbiolit
Questions?
pbourne@ucsd.edu
Recommended