Transcript HCLS Scientific Discourse

HCLS Scientific Discourse
C-SHALS 2009
Tim Clark
Massachusetts General Hospital
& Harvard Medical School
February 25, 2009
Draw conclusions
Draw conclusions
Communicate
Collect data
Collect data
Scientific Discourse
Perform
experiment
Create/modify
hypothesis
Perform
experiment
Gather info
Synthesize
Create/modify
hypothesis
Science is experiment +
discourse
•
•
•
•
Interpretation of experimental evidence;
Placement of results within a model;
Conflicts in models and interpretation;
Resolution of conflicts.
Discourse includes
•
•
•
•
Hypotheses, claims and evidence
Data and interpretations;
Rebuttal, agreement or comment;
Open questions.
Example: Gaps and
Inconsistencies in Research
•
Amongst the 20 top candidate theories
about how Alzheimer Disease works, there
are at least
–
–
•
49 key scientific disagreements and
32 key open questions to be resolved.
Experiment, theory development and
argument should eventually resolve them.
The HCLS Discourse Task
• Develop use cases.
– Neuroscience-related in this case
• Integrate relevant ontologies.
– SIOC, SWAN, myExperiment, Lilly Experiment
• Validate in applications.
• Revise and publish integrated ontologies.
– In progress: SWAN+SIOC IG Note
SWAN+SIOC
• SIOC
http://sioc-project.org
o
Activities and contributions of online communities
o
Integration with blogging, wiki and CMS software
o
Use of existing ontologies e.g. FOAF, SKOS, DC
• SWAN
http://swan.mindinformatics.org
o
Hypotheses, claims, evidence, concepts, entities
o
Schema for SWAN Alzheimer knowledge base
o
Ongoing integration into SCF Drupal toolkit
Digital Enterprise Research Institute
www.deri.ie
SWAN-SIOC integration
• Discourse categories: research questions, scientific
assertions or claims, hypotheses, comments and discussion,
and evidence.
• Biomedical categories: genes, proteins, antibodies, animal
models, laboratory protocols, biological processes, reagents,
disease classifications, user-generated tags, and
bibliographic references.
• Driving Biological Project: cross-application of discoveries,
methods and reagents in stem cell, Alzheimer and Parkinson
disease research.
• Informatics use cases: interoperability of web-based
research communities with
– (a) each other (b) key biomedical ontologies (c ) algorithms for
bibliographic annotation and text mining (d) key resources.
SWAN-SIOC + Experiment
• What the discourse is about, or “warrant
for belief” within discourse.
• Link hypothesis and interpretation to
data and data analysis.
• Meta-analysis, replication, re-analysis.
• Provenance of evidence / analysis.
SWAN-SIOC + myExperiment
DRAFT integration model
Semantic Search Examples
•
•
•
•
•
•
Find all discussion of NF-kappaB by any synonym.
Find discussion on [“NF-KB1” ∧ “inflammatory response”]
Find experiments on other genes and proteins relevant to
“neuroinflammation”.
Find conflicting assertions about PD “anti-inflammatory
response” and supporting experimental data for both.
What workflows were used to analyze the experiments?
Retrieve the data for re-analysis or meta-analysis.
Summary
• HCLS Scientific Discourse connects: hypothesis,
experiment, interpretation, citation, evidence and
discussion.
• Extensive re-use of other ontologies.
• Use case “laboratory” in neuroscience focused on
neurodegenerative diseases and regenerative medicine.
• Progress
– SWAN-SIOC integration complete, IG Note forthcoming.
– SWAN-SIOC myExperiment is in progress.
Credits
•
•
•
•
•
•
•
Uldis Bojars (DERI)
* John Breslin (DERI)
Kei Cheung (Yale)
Paolo Ciccarese (Harvard)
* Tim Clark (Harvard)
Sudeshna Das (Harvard)
David deRoure
(Southampton)
• Ronan Fox (DERI)
• Tudor Groza (DERI)
*task co-chairs
• Christoph Lange (Jacobs
University)
• David Newman
(Southampton)
• Marco Ocana (MGH)
• Alex Passant (DERI)
• Eric Prud’hommeaux (W3C)
• Matthias Samwald (DERI)
• Holger Stenzhorn (DERI)
• Susie Stephens (Eli Lilly)
• Elizabeth Wu (Alzheimer
Research Forum)
With special thanks to
Paolo Ciccarese
David Newman
Alexandre Passant
Eric Prud’hommeaux
Susie Stephens