An Intuitive Graphical Query Interface for Protg Knowledge Bases - - PowerPoint PPT Presentation

An Intuitive Graphical Query Interface for Protégé Knowledge Bases

Landon Todd Detwiler, Cornelius Rosse, Linda Shapiro Structural Informatics Group University of Washington

Emily Emily

Query interface for Prot Query interface for Proté ég gé é KBs KBs

• Graphical, point

Graphical, point-

• and

and-

• click

click

• Stand

Stand-

• alone Java application

alone Java application

• Built on the Prot

Built on the Proté ég gé é API library API library

• Class

Class-

• to

to-

• class relationships

class relationships

Foundational Model of Anatomy (FMA) Foundational Model of Anatomy (FMA) Goal: Usable by novice users with little Goal: Usable by novice users with little instruction instruction

FMA FMA

Foundational Model of Anatomy Foundational Model of Anatomy Symbolic model of the physical Symbolic model of the physical

• rganization of the human body
• rganization of the human body

Goals Goals

• Foundation for interoperability between

Foundation for interoperability between biological domains biological domains

• Basis for the construction of intelligent

Basis for the construction of intelligent biomedical applications biomedical applications

FMA Complications FMA Complications

Space complexity Space complexity

• > 70,000 classes

> 70,000 classes

• > 185,000 frames

> 185,000 frames

• > 170 slots in use

> 170 slots in use

• > 1.5 million slot values

> 1.5 million slot values

Classes without the instances (individuals) Classes without the instances (individuals) Existing tools and the FMA Existing tools and the FMA

Example Questions Example Questions

What is the heart? What is the heart? What are the parts of the heart? What are the parts of the heart? Is the epicardium a part of the heart? Is the epicardium a part of the heart? Which part of the heart is continuous with Which part of the heart is continuous with the superior pulmonary vein? the superior pulmonary vein? What is the relationship between the What is the relationship between the right atrium and blood? right atrium and blood?

Example Questions Example Questions

What is the heart? What is the heart? What are the parts of the heart? What are the parts of the heart? Is the epicardium a part of the heart? Is the epicardium a part of the heart? Which part of the heart is continuous with Which part of the heart is continuous with the superior pulmonary vein? the superior pulmonary vein? What is the relationship between the What is the relationship between the right atrium and blood? right atrium and blood?

Composing a Simple Query Composing a Simple Query

Composing a Simple Query Composing a Simple Query

Subject Object Relationship

What is the heart? What is the heart?

Unknown Unknown has superclass has superclass (directly) (directly) Heart Heart Object Object Relationship Relationship Subject Subject

Subject Tree Subject Tree

Relationship Tree Relationship Tree

Object Tree Object Tree

What is the heart? What is the heart?

Unknown Unknown has superclass has superclass (directly) (directly) Heart Heart Object Object Relationship Relationship Subject Subject Organ with cavitated organ parts Results Results

Example Questions Example Questions

What is the heart? What is the heart? What are the parts of the heart? What are the parts of the heart? Is the epicardium a part of the heart? Is the epicardium a part of the heart? Which part of the heart is continuous with Which part of the heart is continuous with the superior pulmonary vein? the superior pulmonary vein? What is the relationship between the What is the relationship between the right atrium and blood? right atrium and blood?

What are the parts of the heart? What are the parts of the heart?

Unknown Unknown has part has part (directly) (directly) Heart Heart Object Object Relationship Relationship Subject Subject

What are the parts of the heart? What are the parts of the heart?

Example Questions Example Questions

What is the heart? What is the heart? What are the parts of the heart? What are the parts of the heart? Is the epicardium a part of the heart? Is the epicardium a part of the heart? Which part of the heart is continuous with Which part of the heart is continuous with the superior pulmonary vein? the superior pulmonary vein? What is the relationship between the What is the relationship between the right atrium and blood? right atrium and blood?

Is the epicardium a part of Is the epicardium a part of the heart? the heart?

Why was Epicardium not in the previous Why was Epicardium not in the previous result set? result set?

• Epicardium not a direct part of heart

Epicardium not a direct part of heart

• Heart

Heart has part has part Wall of heart Wall of heart which has part which has part Epicardium Epicardium

But aren But aren’ ’t the parts of any direct part of the t the parts of any direct part of the Heart still a part of the Heart? Heart still a part of the Heart?

• Yes, because part is transitive

Yes, because part is transitive

Relationship Tree Relationship Tree

Directly related Transitively related

Is the epicardium a part of Is the epicardium a part of the heart? the heart?

Epicardium Epicardium has part has part Heart Heart Object Object Relationship Relationship Subject Subject Yes Results Results

Example Questions Example Questions

What is the heart? What is the heart? What are the parts of the heart? What are the parts of the heart? Is the epicardium a part of the heart? Is the epicardium a part of the heart? Which part of the heart is continuous with Which part of the heart is continuous with the superior pulmonary vein? the superior pulmonary vein? What is the relationship between the What is the relationship between the right atrium and blood? right atrium and blood?

Which part of the heart is continuous Which part of the heart is continuous with the superior pulmonary vein? with the superior pulmonary vein?

What is the Subject? What is the Subject? What is the Relationship? What is the Relationship? What is the Object? What is the Object?

Compound queries Compound queries

Two methods for creating compound Two methods for creating compound queries queries

• Query chaining

Query chaining

• Set operations

Set operations

Subject: Subject: part of the heart part of the heart Relationship: Relationship: is continuous with is continuous with Object: Object: Superior pulmonary vein Superior pulmonary vein

Which part of the heart is continuous Which part of the heart is continuous with the superior pulmonary vein? with the superior pulmonary vein?

Subject: Subject: part of the heart part of the heart

• Subject:

Subject: Heart Heart

• Relationship:

Relationship: has part has part

• Object:

Object: Unknown Unknown

Relationship: Relationship: is continuous with is continuous with Object: Object: Superior pulmonary vein Superior pulmonary vein

Which part of the heart is continuous Which part of the heart is continuous with the superior pulmonary vein? with the superior pulmonary vein?

U1

Subject: Subject: U1 U1 Relationship: Relationship: is continuous with is continuous with Object: Object: Superior pulmonary vein Superior pulmonary vein

Which part of the heart is continuous Which part of the heart is continuous with the superior pulmonary vein? with the superior pulmonary vein?

Which part of the heart is continuous Which part of the heart is continuous with the superior pulmonary vein? with the superior pulmonary vein?

Which part of the heart is continuous Which part of the heart is continuous with the superior pulmonary vein? with the superior pulmonary vein?

Unknown (U2) Unknown (U2) is continuous with is continuous with U1 U1 Unknown (U1) Unknown (U1) has part has part Heart Heart Object Object Relationship Relationship Subject Subject

Result Sets Result Sets

Query History Panel Query History Panel

U1

Query History Panel Query History Panel

U2

Query History Panel Query History Panel

What part of the heart is What part of the heart is continuous with the superior continuous with the superior pulmonary vein? pulmonary vein?

Compound queries Compound queries

Two methods for creating compound Two methods for creating compound queries queries

• Query chaining

Query chaining

• Set operations

Set operations

Set Operations Set Operations

Results of some queries are sets Results of some queries are sets Available operations Available operations

• AND

AND

• OR

OR

• NOT

NOT

Which part of the heart is continuous Which part of the heart is continuous with the superior pulmonary vein? with the superior pulmonary vein?

Superior Superior pulmonary vein pulmonary vein is continuous with is continuous with Unknown (U2) Unknown (U2) U2 U2 AND AND U1 U1 Unknown (U1) Unknown (U1) has part has part Heart Heart Object Object Relationship Relationship Subject Subject

Set Operations Set Operations

What part of the heart is What part of the heart is continuous with the superior continuous with the superior pulmonary vein? pulmonary vein?

Example Questions Example Questions

What is the heart? What is the heart? What are the parts of the heart? What are the parts of the heart? Is the epicardium a part of the heart? Is the epicardium a part of the heart? Which part of the heart is continuous with Which part of the heart is continuous with the superior pulmonary vein? the superior pulmonary vein? What is the relationship between the What is the relationship between the right atrium and blood? right atrium and blood?

Unknown Relationship Unknown Relationship

What if the user chooses the What if the user chooses the ‘ ‘Unknown Unknown’ ’ node for the query relationship? node for the query relationship?

• FMA is highly connected

FMA is highly connected

Many paths between classes Many paths between classes Not practical (nor useful) to find all paths Not practical (nor useful) to find all paths

• First path

First path

Assumption: The simplest path is likely to be the Assumption: The simplest path is likely to be the most desirable most desirable

Unknown Relationship Unknown Relationship

Algorithm Algorithm

• Direct connections

Direct connections

• Transitive closure connections

Transitive closure connections

• Predetermined connection types

Predetermined connection types

• Depth

Depth-

• limited breadth

limited breadth-

• first search

first search

What is the relationship between What is the relationship between the right atrium and blood the right atrium and blood

Blood Blood Unknown Unknown Right atrium Right atrium Object Object Relationship Relationship Subject Subject Right atrium part Cavity of right atrium contains Blood in right atrium :DIRECT-SUPERCLASSES Blood Results Results

Issues Issues

Translating into S Translating into S-

• R

R-

• O triple

O triple Translating into FMA terms Translating into FMA terms Attributed relationships (e.g. right lateral Attributed relationships (e.g. right lateral adjacency) adjacency)

Conclusion Conclusion

Simple, point Simple, point-

• and

and-

• click query construction

click query construction Query types Query types

• Simple

Simple

• Compound

Compound

“ “Real world Real world” ” limits limits

• Triple formulation

Triple formulation

• Knowledge base content

Knowledge base content