Text mining and Open Access publishing

1
Text mining and Open Access publishing
Matthew CockerillTechnical Director, BioMed
Central
2
Summary

• What is Open Access publishing?
• Open Access publishing and text mining
• About BMC Bioinformatics
• The BioCreative supplement

3
Summary

• What is Open Access publishing?
• Open Access publishing and text mining
• About BMC Bioinformatics
• The BioCreative supplement

4
The current model of publishing scientific
research

• Scientists carry out research
• They write up their results
• They submit them to a journal
• Other scientists act as peer reviewers and
  editorial advisers
• Finally, the publisher sells access to that
  research back to the scientific community

5
Whats wrong with this status quo?

• Restricted access to scientific research is
  contrary to the interests of
• the scientists who do the research
• the funders who pay for it
• society as a whole
• It is an historical artefact of the economics of
  print publishing
• It is a serious obstacle to mining of full text
  information

6
BioMed Central The Open Access publisher

• Commercial organization
• Published first article in mid-2000
• Strict policy of immediate Open Access to all
  research articles

7
Growth of BioMed Central
8
Momentum for Open Access

• PubMed Central
• Public Library of Science
• Open Access declarationsBudapest/Bethesda/Berlin
  
• Software open-source movement
• Mass cancellation of titles from traditional
  publishers

9
BioMed Centrals business model for open access
publishing

• Keep costs down via
• Online submission and peer review
• Automated tools to streamline article processing,
  conversion and layout
• Processing charge (currently 525) for accepted
  articles
• No processing charge for authors at member
  institutions

10
Institutional membership
More than 400 institutions are members of BioMed
Central, including, to name just a few

• CalTech
• Cancer Research UK
• Columbia University
• Cornell University
• University of California
• Dana-Farber Cancer Institute
• Harvard University
• INSERM
• Imperial College
• Institut Pasteur
• John Innes Centre
• Johns Hopkins University
• Kyoto University
• Max Planck Institutes
• Memorial Sloan-Kettering Cancer Center

• MRC Laboratory of Molecular Biology
• National Institutes of Health
• National Institute for Medical Research
• NHS England
• Princeton University
• Rockefeller University
• TIGR
• TSRI
• Tufts University
• Wellcome Trust Sanger Institute
• University of Wisconsin
• World Health Organization
• Yale University

11
Summary

• What is Open Access publishing?
• Open Access publishing and text mining
• About BMC Bioinformatics
• The BioCreative supplement

12
Mining the full text

• Analysing results of high-throughput experiments
  means biologists increasingly need text-mining
  tools
• PubMed is currently the primary resource for text
  mining (its whats available) but
• Abstracts omit critical information
• Techniques developed for abstracts may not
  effectively use extra information in full text
• Fully Open Access corpora, in standard XML
  formats, will help

13
Data mining - BioMed Central
http//www.biomedcentral.com/info/about/datamining

• Entire corpus of full text XML downloadable by
  ftp as a single zip file
• Various groups working with the data
• E.g Pre-BIND (automatic extraction of possible
  protein-protein interaction information from full
  text)
• No restrictions on redistribution
• This means other groups can use same corpus to
  repeat and build on results

14
Data mining - BioMed Central (screen shot)
15
Data mining - PubMed Central
http//www.pubmedcentral.com/about/oai.html

• Standard NLM archiving/interchange XML DTD
  common format across multiple publishers
• Only a subset of PubMed Central participating
  publishers allow download of full text XML
• BioMed Central
• Public Library of Science
• Hopefully, more will follow.
• XML made available via OAI interface

16
Data mining - PubMed Central
17
Adding structure to full text data

• Some examples of useful structure
• Structure of article itself (figure legends,
  materials and methods, references etc)
• MathML, CML etc
• Disambiguated references to genes/proteins

18
Authoring tools are key

• Manuscript structureEndNote, TeX/BibTeX pretty
  good already
• MathML
• Publicon, TeX etc.
• CML
• Chemsketch etc.
• Gene/protein reference markup?
• Semi-automatic markup during authoring
• Author reviews and confirms markup
• System prompts author to clarify ambiguity c.f.
  grammar checker, code intelligence

19
Summary

• What is Open Access publishing?
• Open Access publishing and text mining
• BMC Bioinformatics
• The BioCreative supplement

20
BMC series of online journals

• BMC Biochemistry
• BMC Bioinformatics
• BMC Biotechnology
• BMC Cell Biology
• BMC Chemical Biology
• BMC Developmental Biology
• BMC Ecology
• BMC Evolutionary Biology
• BMC Genetics
• BMC Genomics
• BMC Immunology
• BMC Microbiology
• BMC Molecular Biology
• BMC Neuroscience
• BMC Pharmacology
• BMC Physiology
• BMC Plant Biology
• BMC Structural Biology

• BMC Anesthesiology
• BMC Blood Disorders
• BMC Cancer
• BMC Cardiovascular Disorders
• BMC Clinical Pathology
• BMC Clinical Pharmacology
• BMC Complementary and Alternative Medicine
• BMC Dermatology
• BMC Ear, Nose and Throat Disorders
• BMC Emergency Medicine
• BMC Endocrine Disorders
• BMC Family Practice
• BMC Gastroenterology
• BMC Geriatrics
• BMC Health Services Research
• BMC Infectious Diseases
• BMC International Health and Human Rights
• BMC Medical Education
• BMC Medical Ethics

• BMC Medical Imaging
• BMC Medical Informatics and Decision Making
• BMC Medical Research Methodology
• BMC Musculoskeletal Disorders
• BMC Nephrology
• BMC Neurology
• BMC Nuclear Medicine
• BMC Nursing
• BMC Ophthalmology
• BMC Oral Health
• BMC Palliative Care
• BMC Pediatrics
• BMC Pregnancy and Childbirth
• BMC Psychiatry
• BMC Public Health
• BMC Pulmonary Medicine
• BMC Surgery
• BMC Urology
• BMC Women's Health

21
BMC Bioinformatics
22
RSS feeds
23
Open access leads to high visibility

• Indexing/Linking
• PubMed
• MEDLINE
• ISI
• BIOSIS
• CAS
• CrossRef
• Scirus
• Open Archive Initiative
• Citebase
• Google

• Archiving
• PubMed Central
• INIST
• LOCKSS
• Max Planck
• OhioLINK

24
BMC Bioinformatics - citation impact
25
Summary

• What is Open Access publishing?
• Open Access publishing and text mining
• About BMC Bioinformatics
• The BioCreative supplement

26
Process for publishing in BMC Bioinformatics
supplement

• Follow BMC Bioinformatics Research Article
  instructions for authors
• Send articles to BioCreative organizers who will
  coordinate peer reviewdo not submit articles
  online
• Supplement passed on to BioMed Central for XML
  markup and publication
• 400 processing charge/article

27
Instructions for authors
28
Access to supplement

• All articles in supplement covered by BioMed
  Centrals Open Access licence agreement
• Free access
• Free re-distribution/re-use
• Supplement indexed in PubMed and permanently
  archived in PubMed Central

29
Thats it