XML technologies

1
The ISIS Facilities Ontology and
OntoMaintainerLouisa Casely-Hayford and Shoaib
Sufi
2
ISIS a CCLRC Neutron Muon Facility

• ISIS is the worlds leading pulsed Neutron Muon
  source situated at the CCLRC Rutherford Appleton
  Laboratory. ISIS supports an international
  community of around 1600 scientists in a range of
  scientific disciplines.
• Currently ISIS produces about 700GB of combined
  Neutron Muon data each year and this figure is
  set to rise with the addition of a new target
  station.
• The ISIS Metadata Catalogue (ICAT) is a twenty
  year back catalogue of experiments conducted at
  ISIS it contains approximately 3GB of metadata
  which references 3TB of data.
• In order to maximise the value of data produced
  from the ISIS facility, it must be fully
  searchable.
• To address this problem, e-Science is developing
  numerous software solutions and ontologies are
  seen as one of these useful approaches.

3
Why Ontologies are a useful solution?

• Ontologies offer a powerful means to formally
  express the nature of a domain.
• To share common understanding of the structure of
  information among people
• To enable reuse of domain knowledge
• To make domain assumptions explicit
• They provide central controlled vocabularies that
  can be integrated into catalogues, databases, web
  publications and knowledge management
  applications

4
Why Ontologies?

• Reasons why keywords are limited
• This is because these free text keywords.
• Have no context
• A great deal of their meaning is implicit to ISIS
  users
• Hard to map by non-experts to terms used by
  facilities in the same domain and harder still to
  those outside
• E.g. The keyword HRPD, is a powder
  diffractometer to ISIS users, however other
  collaborating Neutron Facilities like SNS at ORNL
  would understand powder diffractometer but not
  the cryptic HRPD.

• Currently ICAT contains over 10,000 keywords
  describing experiments that are used to index
  experimental studies
• However this is seen as a limited method

5
Why Ontologies are a useful Solution?

• The creation of ontologies at ISIS will aid in
  the mapping of concrete manifestations of
  familiar terms in one domain as well as related
  concepts in different domains.
• This will facilitate searching of data by
  category and grouping of data into keywords
  across studies.
• This could aid in the cross facility searching of
  related scientific data from the various
  scientific facilities housed at CCLRC e.g. CLF
  and DL.

6
Building an Ontology

• Defining terms in a domain and relations between
  them.
• Defining concepts in the domain (classes)
• Arranging the concepts in a hierarchy
  (subclass-superclass hierarchy)
• Defining which attributes and properties (slots)
  classes can have and constraints on their values
• Defining individuals
• Involves collaboration between domain experts and
  ontology builders.
• Ontologies are expressed in a formal language and
  developed within an editing environment.

7
A Protégé-OWL Ontology

• Classes
• Individuals
• Properties

Italy
America
livesIn
Gemma
England
hasSibling
A class is a concept in the domain - a
class of People - a class of Pets - a
class of Countries A class is a collection of
elements with similar properties. Instances of
classes - America can be an instance of the class
Country.
Class Country
Mathew
hasPet
Class Person
Fluffy
Fido
Class Pet
8
Building of the ISIS Facilities Ontology

• Examples of keywords in these five categories
  are
• HRP00145.RAW - a datafile name.
• HRPD - a High Resolution Powder Diffractometer
  one of the many instruments used in experiments
  at the ISIS facility.
• Hydrazinium - an investigation title, chemical
  names and compounds were used as investigation
  titles of experiments in ICAT.
• 1986 - the year in which a particular experiment
  was conducted
• JINR (Joint Institute for Nuclear Research) - the
  name of an investigator.

• The ISIS facilities ontology is based on keywords
  in the ISIS Metadata catalogue (ICAT).
• Over 10,000 keywords housed in ICAT and many are
  synonyms.
• Keywords in ICAT were grouped into 5 main
  categories
• Datafile name
• Instrument
• Investigation title
• Investigator
• Year

9
ISIS Facilities Ontology Hierarchy
10
ISIS Facilities Ontology
Class ISISExperiment
hasTitle
Hydrazinium
Protein Crystallography GroupExperiment
Class InvestigationTitle
wasConductedIn
Class CrystallographyGroupExperiment
1986
hasInvestigator
Class Year
hasUsedInstrument
hasDataFileName
Pete Jones
HRPD
HRP00145.RAW
Class Investigator
Class Instrument
Class DataFile
11
(No Transcript)
12
Further Application of Ontologies to ICAT-ISIS
Online Proposal System

• Scientists can submit applications for beamtime
  at ISIS through an online application form which
  is known as the ISIS Online Proposal System.
• The ICAT(Metadata catalog) not only holds the 20
  year back catalog of data, but will also hold
  data from approved proposals and data generated
  from experiments conducted at ISIS.
• Three separate modular ontologies for Sample,
  Investigator and Experiment are being developed
  to mark up the Proposal system.
• These ontologies are partly based on the proposal
  system database schema.

13
Sample, Investigator and Experiment Ontologies
Sample
Experiment
Investigator
14
OntoMaintainer

• Consensus on Concepts modelled in the ISIS
  Facilities ontology, was achieved through a
  series of interviews with domain experts.
• During the design and creation process, there was
  a difficulty in sharing current versions of the
  ontology with our collaborators at ISIS.
• This is because to view the hierarchical
  structure of the ontology, scientists would have
  to download and install Protégé locally.
• The Ontology Maintainer was developed to
  facilitate the community in remotely viewing
  current versions of the ontology.

15
Screen Shot of OntoMaintainer
16
Benefits of OntoMaintainer

• It is easily accessible because it is available
  over the web
• Allows domain experts to contribute towards the
  maintenance of the ontologies
• Encourages collaboration between domain experts
  (scientists) and ontology builders by allowing
  members of the community to be involved in the
  development and maintenance of ontologies
• Makes collaboration between domain experts and
  ontology builders more efficient

17
Future Work

• Completion of Sample, Investigator, Experiment
  and ISIS Facilities Ontologies
• Ontologies will be used to mark up the ICAT back
  catalogue and new approved studies submitted
  through Online Proposal System
• Ontology Maintainer will be improved through the
  addition of properties to enable relationships
  between individuals in classes to be shown.
• Graphical view of the total hierarchy of the
  ontology will be added to the user interface of
  the Ontology Maintainer.
• Tree hierarchy will be made more dynamic through
  automatic updating of classes
• Currently creating a Mapping tool to map between
  the newly created ontologies and existing
  databases.

18
Conclusion

• Ontologies will help maximise the value of data
  collected at ISIS and other CCLRC facilities by
  improving the access, navigation and reuse of
  data.
• Ontologies will facilitate the mapping of terms
  across CCLRC facilities which will allow
  cross-facility searching e.g. external users will
  be able to search for all experiments carried out
  across CCLRC using a powder diffractometer
  (instrument) even if they do not know the local
  names of the specific instruments.
• The OntoMaintainer will facilitate the process of
  creating and maintaining ontologies by providing
  a means of getting feedback directly from domain
  experts. It will improve the social aspect of
  building ontologies by allowing all members of
  the community to provide input in the building of
  ontologies.
• Major challenge scope, modularity and integration
  of ontologies.

19
References

• Protégé web site http//protege.stanford.edu
• Documentation
• Users Guide
• Tutorial
• Protégé -discussion mailing list
• Ontology library
• CO-ode http//www.co-ode.org/

20
Questions