Transcript Integrating genomic knowledge sources through an anatomy ontology

Integrating genomic
knowledge sources through
an anatomy ontology
Gennari JH, Silberfein A, and Wiley JC
Pac Symp Biocomputing 2005: 115-26
Presented by Morgan Langille
MEDG 505
March 24, 2005
1
Pac Symp Biocomputing?
 Proceedings of the Pacific Symposium on
Biocomputing
 “… research in the theory and application
of computational methods in problems of
biological significance”
 Jan, 2005 @ Hawaii
March 24, 2005
Morgan Langille
2
Outline
Gene Ontology
IntegratedGene Expression
Data
Knowledge
Base!
Foundational Model of
Anatomy
March 24, 2005
Morgan Langille
3
Foundational Model of Anatomy (FMA)
 FMA describes all of human anatomy (even subcellular) as a symbolic ontology of concepts and
relationships
 Designed for the genomics domain not for a
certain type of user
 Can be navigated by humans and machines
 No function or physiology of anatomy
March 24, 2005
Morgan Langille
4
Foundational Model of Anatomy (FMA)
 Implemented using Protégé
 Protégé
 Authoring and editing environment for
ontologies
 Can be used to view the FMA
 Freely accessible since 2003
 Can be viewed in a web browser
March 24, 2005
Morgan Langille
5
March 24, 2005
Morgan Langille
6
Gene Ontology
 Gene Ontology (GO) is a controlled
vocabulary that can be used to annotate
genes
 Includes databases such as:
 FlyBase (Drosophila)
 Saccharomyces Genome Database (SGD)
 Mouse Genome Database (MGD)
 WormBase
 Rat Genome Database (RGD)
March 24, 2005
Morgan Langille
7
Gene Ontology
 The three organizing principles of GO:
 molecular function - catalytic activity,
transporter activity, or binding, etc.
 biological process - cell growth and
maintenance or signal transduction, etc.
 cellular component - rough endoplasmic
reticulum or nucleus, ribosome, proteasome,
etc.
 No tissue specific information
March 24, 2005
Morgan Langille
8
March 24, 2005
Morgan Langille
9
Integrating GO and FMA
 Connect GO with FMA via cellular
structure
 Hand built connections of 150 terms
between FMA and GO
 Built Protégé plugin to view the integrated
data
March 24, 2005
Morgan Langille
10
March 24, 2005
Morgan Langille
11
Gene Expression Data
 No standard source for anatomic knowledge for
annotation of gene expression results
 Standards and Ontologies for Functional
Genomics (SOFG)
 Focused on integrating ontologies for mouse and
human anatomies
 Devoloped “SOFG anatomy entry list” (SAEL)
 SAEL – 100 anatomic terms
 Can be used to annotate gene expression data
March 24, 2005
Morgan Langille
12
Integrating Gene Expression Data
 Integrate the gene expression data from
the Mouse Genome Database (MGD)
 Built connections between anatomy terms
used in MGD to concepts defined in FMA
 Focused only on brain regions
 Few anatomic differences between human
and mouse brain regions
March 24, 2005
Morgan Langille
13
Overview of MGD, GO, and FMA
data integration
March 24, 2005
Morgan Langille
14
Example
March 24, 2005
Morgan Langille
15
March 24, 2005
Morgan Langille
16
Future Work/Problems
 Automation of connecting ontologies
 Prompt
 Plugin for Protégé
 Semi-automatic merging of ontologies
 BioMediator
 Dynamic connections
 Anatomies will not always map between
species
March 24, 2005
Morgan Langille
17
Conclusions
 Many ontologies already exist in biology
such as GO and the FMA
 Integration of multiple sources can be
based on anatomy
 Future work is needed in automating
production of ontology connections
March 24, 2005
Morgan Langille
18
Questions?
 Is anatomy the best knowledge hub?
March 24, 2005
Morgan Langille
19