The Role of Ontologies in Improved Scholarly Communication

1. The Role of Ontologies in Improved Scholarly Communication Philip E. Bourne University of California San Diego pbourne@ucsd.edu http://www.sdsc.edu/pb

2. 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

3. 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

4. First Consider What Motivates Us to Improve Scholarly Communication

5. We Cannot Possibly Read a Fraction of the Papers We Should Drivers of Change Renear & Palmer 2009 Science 325:828-832

6. Hence We Are Scanning More Reading Less Drivers of Change Renear & Palmer 2009 Science 325:828-832

7. 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

8. 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 Computational Biology Resources Lack Persistence and Usability. PLoS Comp. Biol. 4(7): e1000136 Drivers of Change

9. 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 Data and the Publication Are Disjoint Drivers of Change Biosciences Data as of April 14, 2009

10. 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

11. We Need to do Better & The Game is Afoot It is being driven from the top down and the bottom up

12. Ontologies & Semantic Tagging

13. XML XML, Meta-data BioLit Data Extraction/Storage BioLit MySQL database Database IDs Ontology terms Text excerpts Other… <web services> external databases web Semantic Tagging

14. 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

15. Semantic Tagging http://biolit.ucsd.edu

16. ICTP Trieste, December 10, 2007 Semantic Tagging http://biolit.ucsd.edu

17. Provision of Webservices to this tagging may be the most valuable contribution.. Semantic Tagging

18. Database & Literature Integration www.rcsb.org/pdb/explore/literature.do?structureId=1TIM Context Semantic Tagging BMC Bioinformatics 2010 11:220

19. Semantic Tagging of Database Content http://www.pdb.org PLoS Comp. Biol. 6(2) e1000673 Semantic Tagging

20. Cardiac Disease Literature Immunology Literature Automatic Knowledge Discovery for Those with No Time to Read Shared Function Semantic Tagging

21. 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 11:103 Semantic Tagging

22. 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 Drivers of Change http://www.codeplex.com/ucsdbiolit

23. 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 BMC Bioinformatics 2010 11:103 Semantic Tagging

24. 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

25. 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

26. 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

27. Challenges • Author use • Familiarity with ontologies, terms • Agreement between co-authors • End-use of semantically enriched manuscript • Need to combine with NLM XML standard BMC Bioinformatics 2010 11:103 Semantic Tagging

28. Challenges: Author UseIF one or more publishers fast tracked a paper that had semantic markup I would argue it would catch on in no time BMC Bioinformatics 2010 11:103 Semantic Tagging

29. Where we Need {Better} Ontologies1. To Support Mashups Between Different Types of Scholarly Output

30. Post-publication of Video and Paperwww.scivee.tv Drivers of Change

31. Pubcast – Video Integrated with the Full Text of the Paper

32. Pubcasts - A Unique Technology Pubcasts - A Blend of Video, text, tables, figures, PowerPoints, comments, ratings… ALL SYNCHRONIZED FOR RAPID LEARNING 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 Mashups – www.scivee.tv

33. Where we Need {Better} Ontologies2. To Support Tagging of all Aspects of the Scholarly Product

34. Consider Today’s Academic Workflow Reviews Feds Publishers Societies Blogs Community Service/Data Curation Research [Grants] Journal Article Poster Session Conference Paper What Should be Done?

35. Consider Tomorrow’s Academic Workflow Reviews Feds Publishers Societies Blogs Community Service/Data Curation Ideas, Data, Hypotheses Research [Grants] Journal Article Poster Session Conference Paper What Should be Done?

36. Maybe The Line is Somewhere Else? Scientist Laboratory Idea Experiment Data Conclusions Publisher Publish

37. Maybe The Line is Somewhere Else? Laboratory Scientist Idea Experiment Institution Data Lab Notebook Conclusions Publisher Publish What Should We Do?

38. 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

39. 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?

40. 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/

41. 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


42. 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 • LeoChalupa • Lynn Fink • Marc Friedman (CEO) • Ken Liu • Alex Ramos • Willy Suwanto http://www.scivee.tv http://biolit.ucsd.edu http//www.pdb.org http://www.codeplex.com/ucsdbiolit

43. pbourne@ucsd.edu Questions?