Semtech web-protege-tutorial
-
Upload
matthewhorridge -
Category
Education
-
view
3.618 -
download
9
description
Transcript of Semtech web-protege-tutorial
Collaborative Editing of Lightweight Ontologies
withWebProtégé
Natasha NoyMatthew HorridgeTania Tudorache
Stanford University
Download the slides
http://tinyurl.com/semtech-webprotege
Plan
• Introduction
• What is collaborative ontology editing?
• A guided tour of WebProtégé
• Hands-on exercise
• Discussion, Roadmap and Wrap up
What is Protégé?
• An open-source ontology editor• developed at Stanford University• has more than 200,00 registered users• has dozens of plugins for
• visualization• inference• import and export• ….
• has an API for developers
A bit of Protégé history• Started more than 20 years ago
• Has gone through many iterations
• Was the first editor to support OWL 1
• Informed the design of OWL 2
• Has a thriving user community:
• conferences
• mailing list
• short courses
Protégé short course: Vienna, September 2-4
Text
http://protege.stanford.edu/shortcourse/protege-owl/201309/
The “Classic” Protégé
Not what this tutorial is about!
WebProtégé• A Web-based application
• edit ontologies in your Web browser
• nothing to install
• Supports distributed editing
• multiple editors can make changes at the same time
• Includes many collaboration features
• discussion, watches, feeds
Plan
• Introduction
• Collaborative ontology editing
• Hands-on
• WebProtégé in large projects
• Discussion, Roadmap and Wrapup
Collaborative Ontology Development
Collaborative Ontology Development
Collaboration: several users contribute to the
development of one ontology
– Small group → larger community
– Larger ontologies that concern a certain community
– Individual process → social process
Each community does it its own way
Use cases of collaborative development in
biomedical domain
• Gene Ontology (GO)
• NCI Thesaurus
• BiomedGT
• OBI, BIRNLex, RadLex
• Open Biomedical Ontologies (OBO)
• International Classification:
– of Diseases (ICD-11)
– of Traditional Medicine (ICTM)
– of Patient Safety (ICPS)
The NCI Thesaurus collaborative
development process
● Simultaneous editing in Protégé
clients
● Custom UI for restricting user
input and enforcing business
rules
● Development cycle begins after
baseline
● ~20 full-time editors making
changes; 1 “lead editor” who
approves the changes, and
assigns new tasks
● Released version on NCI
website and BioPortal
Reference ontology for cancer biology, translational science, and clinical oncology
ICD-11
● 11th Revision of the International Classification of Diseases
● Over 10.000 categories used for coding, billing, statistics, policy making all over the world
● Collaborative and international effort
● Current version: published as books
● Goal for the new version: use a more formal representation and published in electronical format; use Web-based collaboration and social platforms for editing
Construction of ICD-10: Revision Process in the 20th Century
● 8 Annual Revision Conferences (1982 - 89)
● 17 – 58 Countries participated
– 1- 5 person delegations
– Mainly Health Statisticians
● Manual curation
– List exchange
– Index was done later
● "Decibel” Method of discussion
● Output: Paper Copy
● Work in English only
● Limited testing in the field
ICD-11 process today
● Over 250 domain experts from around the world ● Organized in groups, which edit different parts of the ontology
ICD-11 process today (cont.)
● Each night a snapshot of the commonly edited ontology is published in a public platform to encourage feedback from the larger community http://apps.who.int/classifications/icd11/browse/f/en
● Editorial workflow
● Centrally overseen by WHO
● Peer-reviewed process for the content and structure
● WebProtégé used as the collaborative ontology development platform
Other ways of collaborating: Wikis
● Wikis are well known; Wikipedia
● Semantic Wikis – add semantic extensions to the wiki platforms
● Assign a wiki page to an entity in the ontology (e.g. the class “Mountain”)
● Export/import RDF
Semantic Wiki: MoKi
Source: https://moki.fbk.eu/website/userfiles/image/entmod.png
The challenge with wikis
Source: Hoehndorf, Robert, et al. "BOWiki: an ontology-based wiki for annotation of data and integration of knowledge in biology."
BMC bioinformatics 10.Suppl 5 (2009): S5.
Using sourceforge to manage change proposals for the Gene Ontology
myexperiment: social platform for sharing scientific workflows
Other collaboration processes
● Use source control repositories – SVN, CVS
– Text based mechanisms
– Hard to merge local copies in the shared copy
● Locking mechanisms (lock parts of an ontology for editing)
● Use specialized (domain dependent) ontology repositories, e.g., BioPortal
BioPortal
● An open repository of biomedical ontologies developed by NCBO at Stanford
● Publishing of ontologies, versioning (over 350 ontologies)
● Discussions and structured proposals
● Mappings, views
● Storing metadata
● Search over all ontologies
● Browsing different versions of an ontology
● All content and functionality also available as REST Web services → mash-up of applications
● Technology is domain independent
● http://bioportal.bioontology.org
BioPortal Statistics
Ontology list in BioPortal
NCI Thesaurus details in BioPortal
Useful features for collaboration
● Tools for discussion and reaching consensus
– Add notes to ontology entities (classes, properties, individuals, axioms)
– Add reviews and change proposals anywhere in the ontology
– Document the decision process and final decisions
● Complete Change history
– Establish provenance
– Retrieve ontology snapshots at any time
– Implement different conflict resolution mechanisms
● Personalized views of an ontology based on:
– User’s role and tasks
– User’s level of expertise
Useful features for collaboration (cont.)
● User roles and access control
– Fine-grained control for editing and viewing rights
– Sharing of ontologies
● Publishing released versions of an ontology in a central location,e.g. a repository
● Scalability, reliability and robustness
WebProtégé
A Quick Tour of the UI
Creating an Account I
Create a new account
Creating an Account II
Email address - used for notifications such as ontology changes
User name - displayed next to changes you make and notes that you post
The “Home Screen”
Side bar
Project list. Click projectname to open
Create project
Download project
Sign In/Sign Out
Trash projectUpload project
The Side Bar
All public projects plus your projects that are not in the trash
Your projects that are in the trash
Only projects owned by you that are not in the trash
Projects
A project encompasses: A collection of ontologies
Notes & discussions and watches
Some user interface settings
Some sharing settings
A list of revisions and a log of changes
Creating a ProjectCreate New Project
Project name - does not need to be unique
Project description - appears in the project list
Uploading a ProjectUpload Project
Project name - does not need to be unique
Project description - appears in the project list
Local OWL file name
Sharing
Share link (top right corner)
Public Projects
➊ Select public
➋ Assign permissions for anyone including guests
➌ Assign more fine-grained access for specific users Enter names in list and press “Add”
Private Projects
➊ Select public
Access is restricted to specific users
➋ Assign more permissions for specific users.Enter names in list and press “Add”
Class tree Editor (similar for properties and individuals) Notes & Discussions
Project feed
Editing Class Descriptions
Adding SubclassesCreate subclasses button
Enter one or more names. Press CTRL+Enter to accept and close(one class name per line)
Editing Class Descriptions
Display name - corresponds to the value of rdfs:label here
IRI - Internationalized Resource Identifier. Auto-generated, globally unique
“Property values”(Class expressions under the hood owl:subClassOf)
Annotation assertions
Values can be class names, datatype names, individual names, numbers, dates and strings
Language editor for plain literals
Delete row
Auto-Completion
Type in name. Popup shows possible matches.Dublin Core and SKOS properties “recognised”
On-the-Fly Creation
New property warning(helps prevent typos!)
Press the tab key and enter value to create property(property type will be determined from the value)
Editing Individual DescriptionsClass tree Editor Notes & Discussions
Project feed
Display name - corresponds to the value of rdfs:label here
IRI - Internationalized Resource Identifier. Auto-generated, globally unique
“Property values”(Annotations, property assertions or class expressions under the hood - owl:subClassOf)
Type assertions(rdf:type)
Values can be class names, datatype names, individual names, numbers, dates and strings
Delete row
Same individuals(owl:sameAs)
Editing Individual Descriptions
Icon Cheat SheetClass
Individual (named)
Datatype (xsd:integer, xsd:double etc.)
Property (object/data property)
Annotation property
Number
Date-Time
Literal
Link (http:)
IRI
Hands OnOnline Newspaper
Modelling Task
Build an ontology to describe an online newspaperor news website e.g. www.nyt.com or www.bbc.com
Goal: Become familiar with WebProtégé and some aspects of collaborative ontology editing
Content
Articles:
title, author, date published, edited by, keywords/topics, published in section, media (pictures, video), external links etc.
Advertisements:
Standard ad, personal ad, Service ad etc.
Model different kinds of articles and their properties. For example,
Structure
Newspaper:
date published, issue, front matter etc.
Sections:
Domestic News, World News, Editorial, Magazine, Letters, Commentary, Television Listings, Advertisements, Appointments/Jobs, Sport, Business etc.
Sections and subsections
Model the structure of a news paper - different sections and how theyfit together. For example,
People
Employees:
Columnist, Editor, Section Editor, Reporter, International Reporter, Manager
name, contact details: email, phone number, role
Other people:
Politician, President, Actor etc. Individual people, e.g. Barack Obama.
Model the people who contribute to the news paper and people whoare the subject of articles. For example,
Custom entry forms for editing the ontology content
● Easy to create user interfaces for the domain experts
● Use common entry forms, but still keep the ontology “intelligence” behind it
● A form widget (e.g., text field) is linked to a property in the ontology
● Easy to create custom forms with different views for different users
● Hides complex ontology stuff
Form configuration in WebProtégé
Form-based editing and configuration of the user interface for the development of ICD-11
http://icatdemo.stanford.edu
Examples of form-based editing
Importing BioPortal terms into WebProtégé
(1) Search term in BioPortal ontologies
(2) Get
search
results(3) Browse
details of
results
(4) Import into WebProtégé with
single click
WebProtégé – Make Up
Protégé Collaboration
Framework
WebProtégé
WebProtégé Server
GWT RPC
Server side
Client side
Java
Java
Java at
development time
JavaScript at
run- time
2 parts: server and client
Server is completely
implemented in Java and makes
API calls to the OWL-API and
other libraries
Client side is developed in Java,
and later compiled by GWT into
JavaScript
Communication between server
client is done via GWT RPC or
simple HTTP calls
WebProtégé is pluggable
WebProtégé User Interface (GWT)
Portlets
Event manager Other managers
WebProtégé Server (Java)
Access policies service
...Ontology Service
Notes and changes Service
pluggable
pluggable
Extending WebProtégé
Plug-in infrastructure very similar to Protégé's: create your
own tabs and portlets
Extend: AbstractTab or AbstractEntityPortlet
Implement your own RPCs, if needed
Reuse existing portlet code
Writing a tab – as easy as creating an empty class that
extends AbstractTab
http://protegewiki.stanford.edu/wiki/WebProtegeImplementationGuide
Resources
● Online WebProtégé server: http://webprotege.stanford.edu
● WebProtégé documentation:http://protegewiki.stanford.edu/wiki/WebProtege
● WebProtégé paper: “WebProtégé: A Collaborative Ontology Editor and Knowledge
Acquisition Tool for the Web”, Tania Tudorache, Csongor Nyulas, Natalya F. Noy,
Mark A. Musen, Semantic Web Journal (SWJ) 4 (Number 1 / 2013), 89 - 99
● WebProtégé in use: “Will Semantic Web Technologies Work for the Development of
ICD-11?”, T. Tudorache, S. M. Falconer, C. I. Nyulas, N. F. Noy, M. A. Musen. The 9th
International Semantic Web Conference, ISWC 2010 (In-Use track), Shanghai,
China, Springer. Published in 2010.
http://bmir.stanford.edu/file_asset/index.php/1646/BMIR-2010-1427.pdf
● Other References: http://protegewiki.stanford.edu/wiki/WebProtege#References