Investigational Ontology Requirements for Neuroinformatics

1
Investigational Ontology Requirements for
Neuroinformatics
1st OBI Meeting - 01/29/2007
2
Neuroinformatics

• Definition
• Computational representation, manipulation,
  analysis of experimental evidence captured during
  neuroscientific investigation
• Representational Domains
• Formal annotations of data - spatially-localizatio
  n
• Literature Informatics - IR --gt NLP
• CocoMac, BrainMap.org, Neuroscholar/NeuArt,
  Arrowsmith
• Models
• molecular, cellular, and systems - function
  structure
• Distilled Knowledge Resources
• Minimal KM NeuroNames,, SenseLab Terms, NIF
  terminology, IUPHAR receptor/channel, PDSP Ki,
  etc.
• Ontologies FMA, CL, SAO, BIRNLex

3
Neuroinformatics

• Long-term Goal
• Promote neuro-wide, contextually relevant data
  pooling for large-scale, multi-rez,
  multi-modality meta-analysis
• Requires formal representation of 3Es
• Evidence (PATO)
• Experimental context (OBI)
• Enviromental factors (EnvironmentO)

4
Neuroinformatics

• OBI Focus
• Community shared semantics for wide-variety of
  investigational techniques
• Investigational Domains
• Molecular (EM,molecular physiology, FISH, IHC,
  receptor binding)
• Cells Cell Components (EM, LM, histo, cell.
  assays )
• Neuroanatomy (microscopy radiological imaging)
• Physiology (cellular system physiology)
• Behavior Cognitive (assessments)

5
Neuroinformatics

• All applies to single lab studies analyzed in
  isolation
• Large-scale neuroinformatics
• Lit. informatics
• BIRN
• NIF
• NITRC
• NeuroBase
• SenseLab
• MBL/GeneNetwork
• CCDB
• Atlases - LONI MAP, NT, Hof et al, Mouse
  ImgCenter (Toronto)
• Brain image data sets - GENSAT, ABA, Smart Atlas,
  BrainMaps.org

6
BIRN Ontology Use
Critical to BIRN-wide

• semantically-mediated data integration

• automated data validity integrity

• data pooling/binning/sorting/modes

• data reduction

• analysis

• knowledge extraction, representation, discovery
  (KE/KR/KD)

Integration across BIRN sites

• 29 locations - multiple labs/site

• with the rest of the world

Supports access both by human experts algorithms
7
BIRN Ontology Task Force (BIRN OTF)
Goals

• Develop ontology usage policies and procedures
  for BIRN participants

• Biomedical ontology development community outreach

• Monitor community activity, participate
  disseminate

• Educate BIRN participants on the use, development
  and importance of ontologies

• Promote and develop cross test bed ontology
  frameworks to support KE/KR/KD for animal models
  and human disease

8
BIRN OTF
Personnel
BIRN-wide Representation

• BIRN-CC

• Jeff Grethe (UCSD)

BIRN-CC Scientific Coord. co-Inv.

• Amarnath Gupta (UCSD)

BIRN-CC co-Inv.- BIRN Mediator

• MBIRN

• Christine Fennema-Notestine (UCSD)

• David Kennedy (MGH)

• FBIRN

• Jessica Turner (UCI)

• mBIRN

• Maryann Martone (UCSD)

• Bill Bug (Drexel U.)

• NIH NCRR/NCBO

co-chair

• Carol Bean (NIH)

honorary

• Daniel Rubin (SMI, Stanford U.)

9
BIRN OTF
Work-in-progress

• BIRN IT Infrastructure maturing

• Intra-BIRN and external collaborative practice
  maturing

• Broad spectrum of neuroimage-related tools data

• software APIs for handling files - many
  image-specific

• spatially-mapped neuro-data sets

• visualization, spatial normalization analysis

• workflow - models of computation (e.g., Kepler,
  OpenWFE)

• BIRN ontology use

• cross modality techniques, curatorial metadata,
  harvest low KM terms stay in sync, image
  annotation, cross-species, OWLMediation

10
BIRN OTF
Best Practices

• if ontology exists for required knowledge domain

• re-use it

• biomedical KE/KR/KD has developed considerable
  lexical ontological sophistication over the
  last decade

• re-use formalisms tools as well (e.g., OWL,
  SKOS, PaTO-OBI, etc.)

• unfortunately, some existing ontologies

• dont exist yet for all neuroscience domains -
  e.g., physiology, cognition

• conflate/pre-coordinate entities requiring
  cleansing beyond our definition of re-use

• cover required domains but in a manner not
  fitting a neuroscience context

• use incompatible formalisms

• Need for neuroinformatic outreach

• Many of the caveats above are being addressed in
  last few years - our participation would help
  hasten this effort in a way useful to
  large-scale, field-wide semantically-driven
  neuroinformatic query systems

11
BIRN OTF
Best Practices

• Need to start semantic annotation in BIRN NOW -
  therefore

• systematically collect research lexicon in use by
  40 BIRN labs

• re-use ontologies in current form if feasible -
  even if not ontologically ideal

• represent formally using Protégé-OWL

• follow community best-practices, and try to be
  formally precise in constructing the semantic
  graph so as to promote clarity and lexical
  separation

• use specific curatorial typing (i.e., BIRN
  curation status tags)

• use normative lexical variant typing (e.g., SKOS)

• Should hasten effort to ultimately translate
  BIRNLex into a collection of formal BIRN
  ontologies

• The multi-scale investigation of neurological
  disease ontology framework

12
BIRN OTF
Knowledge resource development

• BIRN Project - comprehensive Use Case set

• Excellent test bed for research use of biomedical
  experiment ontology integrated with other
  ontologies in the biomedical domain

• BIRN OTF gathers, collates, publishes Use Cases
  across test beds

• Ontology tools ontology use rapidly evolving

• BONFIRE (tool)

• BIRNLex (lexicon)

• MIND (ontology) - only what we cannot re-use

• BIRN ontology best practices - social
  engineering

13
BIRN OTF
Work-in-progress

• How to actually link to use external ontologies

• when and how to resolve URIs

• BIRN Investigational Ontology elements

• device, assay, reagent, protocol

• thinnest component of BIRNLex now - granularity
  non-uniform

• began with FuGO/OBI from March 2006

• some additional curation required to cleanse
  current experiment entities - shows UMLS origins

• still hope to get most of what we require from
  OBI, PaTO, ChEBI, GO, CL-O, Neuro-FMA, etc.

14
Mouse BIRN
Experiment Ontology Requirements
Use Cases illustrating experiment ontology
requirements
15
Brain Atlas System Integration
BIRN Atlas Interoperability API
SMART Atlas (UCSD)
NeuroTerrain (Drexel)
SHIVA (UCLA)
BIRN Atlas Interoperability API
MBAT (BIRN)
16
(No Transcript)
17
(No Transcript)