Introduction to Ontologies and the AmphibAnat Project

1
Introduction to Ontologies and the AmphibAnat
Project

• 2007 AmphibAnat Meeting
• Wed, Nov 14 900930
• Anne Maglia

2
Project history

• AmphibiaTree workshop Austin, 2004
• An anatomical ontology is necessary in order to
  facilitate morphological character coding,
  especially across larger groups
• AmphibiaTree workshop Lawrence, 2006
• An anatomical ontology would allow the
  integration of genomic, developmental, and
  morphological data, and help standardize
  phylogenetic characters and morphological
  descriptions

3
AmphibAnat

• Semi-automated construction of an ontology for
  amphibian morphology
• Jennifer Leopold (PI), Anne Maglia, Analía
  Pugener, and Susan Gauch
• NSF DBI-0640053 1,116,729, 36 mo.
• Participant costs annual workshop/board meeting
• Cost of development and maintenance of the
  ontology, ontology-building software, database,
  and user interface

4
Outline

• What is an ontology?
• Other closely-related projects
• Approach to developing AmphibAnat
• What we hope to cover and accomplish today

5
In philosophy

• The study of being or existence
• Describes the basic categories and relationships
  of being to define entities and types of entities
• Study of the conceptions of reality
• Nouns are entities or collections of objects or
  events

6
In computer science

• A data model that represents a domain, the
  objects in that domain, and the relations between
  them
• A form of knowledge representation about the
  world or some part of it

7
Informally

• A set of definitions of formal vocabulary
• Includes an agreement to use a vocabulary in ways
  that are consistent (but not complete) with
  respect to the theory specified by the ontology

8
A working definition

• A formal vocabulary, in the form of a knowledge
  representation, that describes the basic
  categories of being by defining entities, types
  of entities, and the relationships among them

9
Elements of an ontology

• Instances (individuals)
• Classes (concepts)
• Attributes (properties)
• Relationships
• Hierarchical structure

10
Instances

• Also called individuals
• The most basic components of an ontology
• The actual, concrete objects
• An ontology does not require the inclusion of
  instances, but a main purpose of an ontology is
  to provide a means of classifying individuals,
  even if those instances are not explicitly part
  of the ontology

11
Examples of instances

• Instance David Cannatella
• an instance of Homo sapiens
• Instance KU103854
• an instance of Dermophis mexicanus
• Instance UT-Austin
• an instance of a university campus
• Instance Maglias car
• An instance of a Toyota Prius

12
Classes

• Also called concepts
• Abstract groups, sets, or collections of objects
• May contain individuals, other classes, or a
  combination of both

13
Examples of Classes

• Class Vehicle
• Class Car
• Class Toyota
• Class Hybrid
• Class Prius

14
Attributes

• Also called properties
• Features, characteristics, or parameters that
  objects can have and share
• Objects in the ontology are described by
  assigning attributes to them

15
Example of attributes

• Class Prius
• Attribute Number_of_doors Value four
• Attribute Engine Value hybrid synergy
• Instance Maglias car
• Attribute Name Value Maglias car
• Attribute Color Value Silver
• Attribute Registration_state Value Missouri

16
Relationships

• Ways that objects interact with one another
• An attribute whose value is another object in the
  ontology
• Most important type is subsumption is_a

17
Examples of relationships

• Instance Maglias car
• Relationship is_a Class Prius
• Relationship is_same_color_as Instance
  Leopolds car
• Class Prius
• Relationship is_a Class Toyota
• Relationship is_smaller_than Class 4-Runner

18
Hierarchical structure

• Inherent in classification system
• Defined by relationships among classes
• Attributes of superclasses are inherited by
  subclasses
• Most commonly used are is_a, part_of, and
  develops_from

19
Is_a classification

• Class Vehicle
• Class Truck
• Class Car
• Class Honda
• Class Toyota
• Class Gasoline
• Class Hybrid
• Class Prius

20
Part_of classification

• Class Vehicle
• Class Frame
• Class Engine
• Class Electrical system
• Class Transmission
• Class Sprockets
• Class Bearings

21
Other ways to represent knowledge

• Controlled Vocabulary classes only, no
  attributes, no relationships, no hierarchy
• Anatomical Dictionary classes and attributes, no
  hierarchy, no relationships
• Taxonomy classes, hierarchy, and relationships,
  no attributes

22
Projects close to AmphibAnat

• Xenopus anatomy ontology

23
Projects close to AmphibAnat

• Zebrafish anatomical ontology

Cypriniformes and Teleost anatomical ontologies
24
AmphibAnat
Literature citations
Development ontology
Anatomical ontology
Images
Phenotype
Taxonomy
User comments
25
Amphibian Anatomical Ontology

• Semi-automated construction approach
• Develop IE software that reads literature to
  add concepts and relationships
• Combine with user-community expertise
• Manually-building seed ontologies
• Augmenting, evaluating, and refining
  software-generated ontologies
• Alternative approach to ontology maintenance
• Comprehensive functionality
• Necessary functions

26
Goals for this workshop

• Introduce/reintroduce AmphibAnat and ontologies
• Become familiar with AmphibAnat methods of
  construction and augmentation/community
  acceptance
• Evaluate website and RDBOM interface
• Provide usability feedback
• Enhance expert user knowledge of, interest in,
  and participation in project