Supporting Systematic Ontology Creation and Reuse with OntoGuide - PowerPoint PPT Presentation

Title: Supporting Systematic Ontology Creation and Reuse with OntoGuide


1
Supporting Systematic Ontology Creation and Reuse
with OntoGuide

• UFRJ (Federal University of Rio de Janeiro -
  Brazil)
• NCE - 2004

Linair Maria Campos, MSc StudentMaria Luiza M.
Campos, PhD Paulo F. Pires, DSc
2
Summary

• Genomic data and Ontologies
• The Problems of Ontologies Building
• Objective
• Existing Approaches
• Proposed Approach
• Conclusion

3
Genomic data and Ontologies

• The growth of genome data and genome analysis
  initiatives has brought new levels of difficulty
  for scientists to understand, integrate and deal
  with all this ever-coming information -
  ambiguity is frequent when describing those data,
  which leads to the need of a consensual solution.
  
• ?

Ontologies come as an attractive choice to reach
consensus, as they offer standard terminology and
reasoning mechanisms to represent and infer new
knowledge . However
4
The Problems of Building Ontologies

• There is no unanimous way to define and organize
  concepts and their relations in a systematic way
• How to define a concept ?
• How to define relations?
• How to reuse an ontology?
• ?

ltltour case study focuses in finding
Bioinformatics Web services, but can be applied
to any domaingtgtThis is the kind of question we
wish to answer Which are the free
Bioinformatics Web Services that align nucleotide
sequences against genomic databases, using Blast
algorithm, with average performance of 1.0 ms,
using GenBank, available Sunday?
5
The Problems of Building OntologiesWeb Services
case DReggie X OWL-S
Fine-grained X Corse-grained
?
6
The Problems of Building Ontologies

• How to better organize the attributes of a given
  domain?
• How to facilitate the extension and creation of
  vocabularies needed to describe that domain?
• How to stimulate the reuse of existing
  vocabularies?

7
Objective

• To present a systematic approach for organization
  and definition of a set of domains descriptors
• considering reuse
• to help retrieve domains objects based on their
  diverse aspects
• supported by a tool - OntoGuide

8
Proposed Systematic Approach Steps

• We looked for help from Information Science and
  Terminology Areas mature experiences
• Requisites for elaboration of conceptual model of
  hyperdocuments
• Thorough study of information organization
  (Concept Theory, Faceted Classification,
  Terminology )
• (M.L.A.Campos proposal)

• Scope Definition and Domain Organization
• KU Definition (concepts) and their Relations
• KU detailed description
• Revision of Definitions (cyclic process)

9
Proposed Approach Step 1

• Scope Definition and Domain Organization

?
10
Proposed Approach Step 2

• KU Definition and KU Relations
• A method is needed

• The Concept is represented by a term whose
  meaning is context dependent.
• Concepts characteristics should be identified
  and expressed in sentences.
• Common characteristics are used to group
  concepts in classes.
• Relevant characteristics should be present on
  the concepts definition.

?
11
Proposed Approach Step 2

• When defining a concept be sure to establish
• What is it
• Which are their parts (if applicable)
• For what it is used ?

Ex A restriction-enzyme is a enzyme it is
composed of aminoacids it is used for breaking
DNA molecules
12
Proposed Approach Steps 3 / 4

• KU details

Through the whole processDocument each
decision!E.g. we use Ranganathans Fundamental
Categoriesand categories of relations used on
Terminology
?
13
OntoGuide

• We have implemented our directives on a tool
  named OntoGuide
• In OntoGuide we deal with OWL ontologies
• 1 Configuring (by choosing ontologies
  Fundamental categories)
• 2 Treating (by associating Fundamental
  categories)
• 3 Extending (by reusing existing ontologies)

14
OntoGuide Configuration
?
15
OntoGuide Configuration existing ontologies
?
16
OntoGuide Ontology Extension
?
17
OntoGuide Ontology Extension
?
18
Tests and Results
Our approach is flexibleAll services were found
?

• We look for a set of Bioinformatics Web services
  to
• Search for genes
• Align nucleotide seq. returning the best matching
• Align nucleotides globally
• Elaborate a phylogenetic tree, using maximum
  parsimony

19
Conclusion Rescuing the Problem

• The crescent need of complex ontologies has
  created challenges to their elaboration
• Existing initiatives are too generic
• Easy extension but non-uniform vocabularies
• Or are too specific
• Do not solve many problems
• They dont consider descriptors organization
• Our approach tackles those issues

20
Conclusion - Contributions

• A set of directives that provide a systematic
  approach towards creation, extension and
  organization of domains descriptors
• A tool that applies those directives
• An automated way of documenting vocabularies.

21
Conclusion - Contributions

• The use of a multidisciplinary approach
• Bringing to Computer Science points of view from
  Information Science and Terminology may be used
  in future work
• Applying in Bioinformatics, using in partnership
  projects in this area at Fiocruz Foundation

22
Conclusion Limitations Difficulties

• Few ontologies available for use or with scarce
  documentation
• Among them, many are appropriated for learning
  purposes, but not for effective usage.
• Simple comparative analysis due to the lack of
  real examples.

23
Conclusion Future Work

• Explore the use of different approaches that
  apply enhanced formalism on concepts definition
• Doctoral thesis of Simone Garcia at UFRJ
• Use OntoGuide as a Protégé Plugin

24
Supporting Systematic Ontology Creation and Reuse
with OntoGuide

• UFRJ (Federal University of Rio de Janeiro -
  Brazil)
• NCE - 2004

http//dataware.nce.ufrj.br
Linair Maria Campos, MSc Student linairmc_at_posgrad.
nce.ufrj.brMaria Luiza M. Campos, PhD Paulo F.
Pires, DSc