Top Banner
cBio Meeting, March 2-3, 2006 CHISEL Group Dept of Computer Science University of Victoria, Canada Visualization of ontologies and data annotations
24

CHISEL Group Dept of Computer Science University of Victoria, Canada

Jan 19, 2016

Download

Documents

steffi

Visualization of ontologies and data annotations. CHISEL Group Dept of Computer Science University of Victoria, Canada. Team Members. Margaret-Anne Storey (PI) Chris Callendar (Programmer) Visualization framework for Bioportal Degree of Interest model implementation - PowerPoint PPT Presentation
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: CHISEL Group Dept of Computer Science University of Victoria, Canada

cBio Meeting, March 2-3, 2006

CHISEL GroupDept of Computer ScienceUniversity of Victoria, Canada

Visualization of ontologies and data annotations

Page 2: CHISEL Group Dept of Computer Science University of Victoria, Canada

cBio Meeting, March 2-3, 2006

Team Members

• Margaret-Anne Storey (PI)

• Chris Callendar (Programmer)– Visualization framework for Bioportal– Degree of Interest model implementation

• Tricia d’Entremont (PhD student)– Pictorial based ontology navigation (assisting Nigam)– Degree of interest models and ontologies

• Sean Falconer (PhD student)– Ontology search – Algorithms and visualizations of ontology alignments

• Maria-Elena Hernandez (Phd student)– Visualizing data annotations – clinical trial

visualization

Page 3: CHISEL Group Dept of Computer Science University of Victoria, Canada

cBio Meeting, March 2-3, 2006

Our research goals

Core 1: • Develop visualization services for Bioportal

– A visualization toolkit– A visualization ontology– A mapping mechanism to specify how to integrate

and customize the available services for particular ontologies and tasks

• Evaluate through – instrumentation and integration case studies

Core 2:• Visualization of data annotations and meta-

data analysis– HIV clinical trials– Phenotype annotations

Page 4: CHISEL Group Dept of Computer Science University of Victoria, Canada

cBio Meeting, March 2-3, 2006

Approach

• Determine requirements for visualization in an iterative manner: – Identify different user groups and user

tasks (driven by the Core 3 projects)

• Draw from research on:– Human computer interaction– Visualization– Adaptive interfaces– Computer supported collaborative work

Page 5: CHISEL Group Dept of Computer Science University of Victoria, Canada

cBio Meeting, March 2-3, 2006

Previous work and expertise

• Jambalaya – visualization support for Protégé

• Developed and evaluated visualization tools to support comprehension, navigation and collaboration in software engineering

Page 6: CHISEL Group Dept of Computer Science University of Victoria, Canada

cBio Meeting, March 2-3, 2006

History of Jambalaya

• What is Jambalaya?– Jambalaya = SHriMP + Protégé & Protégé-

OWL glue– What is SHriMP?

• Nested (or un-nested) graph• Smooth animated zooming & graph layouts• Embedding of AWT/Swing widgets (e.g.

Protégé forms) within visualization

Page 7: CHISEL Group Dept of Computer Science University of Victoria, Canada

cBio Meeting, March 2-3, 2006

Jambalaya: Protégé + SHriMP

Page 8: CHISEL Group Dept of Computer Science University of Victoria, Canada

cBio Meeting, March 2-3, 2006

Express views

“Express View” buttons

“Root Classes”

Page 9: CHISEL Group Dept of Computer Science University of Victoria, Canada

cBio Meeting, March 2-3, 2006

Query View

• Bottom up approach instead of top down (like others)

• invoked from Class’ right-click menu, on any tab

Page 10: CHISEL Group Dept of Computer Science University of Victoria, Canada

cBio Meeting, March 2-3, 2006

PromptViz: visualizing two versions of an ontology

Page 11: CHISEL Group Dept of Computer Science University of Victoria, Canada

cBio Meeting, March 2-3, 2006

Determining requirements

• Unanswered questions…– Where should a visualization be used?– When would the user want to see it?– How is the data best represented in the visualization?

• Fundamental Visualization Question– How can we provide a useful visualization at the

moment a user needs it?

• Need better support for task-driven “visualization-on-demand”– Visualizations should be readily available from familiar

tools– A visualization should immediately answer a specific

question or support a particular task– They should not take too much effort to generate and

they should be efficient

Page 12: CHISEL Group Dept of Computer Science University of Victoria, Canada

cBio Meeting, March 2-3, 2006

Views currently proposed for the toolkit

• Overviews• Hub concepts • Query views• Navigational views• Difference views

Page 13: CHISEL Group Dept of Computer Science University of Victoria, Canada

cBio Meeting, March 2-3, 2006

Engineering goals and Evaluation

Services provided by the toolkit will have well-defined and clearly documented APIs to support the BioPortal tools

• Software engineering principles:– Interoperability– Customizability– Extensibility

• Evaluation:– Instrumentation– Integration case studies

Page 14: CHISEL Group Dept of Computer Science University of Victoria, Canada

cBio Meeting, March 2-3, 2006

Demonstrations of Current Work

• Degree of Interest Model (Chris)• Jambalaya Lite Applets (Chris)

• Ontology Search (Sean)

• Pictorial based ontology navigation (Nigam/Tricia)

Page 15: CHISEL Group Dept of Computer Science University of Victoria, Canada

cBio Meeting, March 2-3, 2006

Degree of Interest Model

• To address the problem of information overload with large ontologies and to identify relevant information

• A degree of interest model is developed by monitoring the user’s activities (e.g. navigation actions, editing and annotations)

• The model is used to highlight or filter more “interesting” elements in the ontology

• Extending the work of Stuart Card (Degree of Interest Trees) and Mik Kersten (Mylar)

Page 16: CHISEL Group Dept of Computer Science University of Victoria, Canada

cBio Meeting, March 2-3, 2006

Jambalaya Lite with Mylar activated

Page 17: CHISEL Group Dept of Computer Science University of Victoria, Canada

cBio Meeting, March 2-3, 2006

Page 18: CHISEL Group Dept of Computer Science University of Victoria, Canada

cBio Meeting, March 2-3, 2006

Page 19: CHISEL Group Dept of Computer Science University of Victoria, Canada

cBio Meeting, March 2-3, 2006

Page 20: CHISEL Group Dept of Computer Science University of Victoria, Canada

cBio Meeting, March 2-3, 2006

Ontology Alignment

• Current approaches – rely heavily on syntactic comparison– problems with synonymous concepts

• Need semantic comparison– skull -> cranium– DNA -> deoxyribonucleic acid

Page 21: CHISEL Group Dept of Computer Science University of Victoria, Canada

cBio Meeting, March 2-3, 2006

Ontology Alignment

• Developed semantic comparison algorithm– Can match synonyms, abbreviations,

phrases, etc.

• Tested against synonym datasets• Plan to develop a new alignment

algorithm and incorporate this technique

Page 22: CHISEL Group Dept of Computer Science University of Victoria, Canada

cBio Meeting, March 2-3, 2006

Pictorial guided ontology navigation

Page 23: CHISEL Group Dept of Computer Science University of Victoria, Canada

cBio Meeting, March 2-3, 2006

Short term goals (next year)

• Jambalaya Lite applets – deploy and integrate with Bioportal (May 2006)

• Initial prototype of Jambalaya thin client for Bioportal (Dec 2006)

• Degree of interest model – evaluate through integrations with Protégé and OBOEdit (Dec 2006)

• Ontology search (September 2006)• Support for visualizing clinical trial data (Dec

2006)

Page 24: CHISEL Group Dept of Computer Science University of Victoria, Canada

cBio Meeting, March 2-3, 2006

Longer term goals

• Visualization framework: [Chris]– Toolkit of visualization views and widgets– Visualization ontology– Mappings (graph transformations)– Visualization thin client for Bioportal:

proposed technology SVG and Ajax

• Ontology alignment visualizations and algorithms [Sean]

• Degree of interest model integrated and evaluated with Bioportal [Tricia]

• Visualizing data annotations across the DBPs [Maleh]