Docstoc

BAO Journal Club

Document Sample
BAO Journal Club Powered By Docstoc
					BAO Journal Club


 December 7th, 2009
    Robin Smith
Pubchem assays by date
2500
           @Time of grant proposal: 1500
           @Time of Robin start: 1885
           @Yesterday: 1957                                                                         R² = 0.9635
2000



1500



1000



 500



   0
   2/17/2005   9/5/2005   3/24/2006   10/10/2006   4/28/2007   11/14/2007   6/1/2008   12/18/2008   7/6/2009      1/22/2010
Bioassay Ontology: top down approach
• How best to represent high
  throughput/content experiments using OWL?
• Distinguish between simple concepts, such as:
  – How? How was the assay carried out. This
    includes the underlying technology.
  – What? What was measured by the experiment?
    A change in kinase activity or an increase in cell
    number?
  – Which? Which biological system was the assay
    carried out in? A reconstitued complex, a cell line
Bioassay Ontology: key concepts
• AssayBiologicalSystem
  – The environment in which the bioassay is being
    conducted, in addition to a specific target information.
• AssayMeasure:
  – The physical process or activity that is being measured
    by the bioassay.
• AssayTechnology
  – The underlying techniques used to obtain the results
    of the assay.
• AssayDesign
  – General features of the assay, including the nature of
    the inputs and endpoints, as well as the assay’s status
    within a larger campaign.
Is top-down ontology design best?
• What’s the best way to build an ontology?


    Ph.D.
Domain specific vocabulary extraction

                             Domain Specific Vocabulary

                                     concentration
                                 Cell-based
                              luciferase
                                     promoter
                                           Enzyme

                                                Ph.D.
NCBO Bioportal Annotator
• ~200 biomedical ontologies
• ~1.5 million concepts
• Can we extract domain specific vocabulary to
  aid in ontology construction?
  – Query web service using Pipeline Pilot
The “SNOMED” (and NCI) problem




        50,000 + annotations for each of name, description and protocol
NCBO Recommender service
Some ontologies may be useful
Domain expert vocabulary extraction

                            Domain Specific Vocabulary
                   Ph.D.            concentration
                                Cell-based
                             luciferase
                                    promoter
                                          Enzyme
PubChem – Spotfire representation
Simple AssayFormat Annotation


           216

                          Biochemical

                 732
                          CellBased


                          Unsure
                          (Summary/Other)
        1009
Can use high level divisions to find smaller
assay groups: has “luciferase”?
Can use high level divisions to find smaller
assay groups: has “enzyme”?
Pubchem assays by screening stage
                        136
                137

                                              confirmatory
                                   819        screening
          301
                                              other
                                              summary
                                              null

                      493



 • Screening assays and some “other” assays are difficult to annotate
 • Some “other” assays are in vivo screens
 • “Null” = on hold

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:4
posted:7/22/2012
language:
pages:16