W3C HCLS Dataset Description Guidelines

Describing Scientific Datasets:
The HCLS Community Profile
1
Michel Dumontier, Ph.D.
Associate Professor of Medicine (Biomedical Informatics)
Stanford University

World Wide Web Consortium (W3C)
• The W3C is the main international standards
organization for the World Wide Web.
• The W3C is made up of over 400 member
organizations for the purpose of working
together in the development of standards for
the World Wide Web.
@micheldumontier::CEDAR:Jan 20152

The Semantic Web
is the new global web of knowledge
3 @micheldumontier::CEDAR:Jan 2015
It involves standards for publishing, sharing and querying
facts, expert knowledge and services
It is a scalable approach to the
discovery of independently formulated
and distributed knowledge

Resource Description Framework
• It’s a language to represent knowledge
– Logic-based formalism -> automated reasoning
– graph-like properties -> data analysis
• Good for
– Describing in terms of type, attributes, relations
– Integrating data from different sources
– Sharing the data (W3C standard)
– Reusing what is available, developing what you need,
and contributing back to the web of data.

drugbank:DB00586
drugbank_vocabulary:Drug
rdf:type
drugbank:290
drugbank_vocabulary:Target
rdf:type
drugbank_vocabulary:targets
rdfs:label
Prostaglandin G/H synthase 2
[drugbank_target:290]
rdfs:label
Diclofenac [drugbank:DB00586]
5
PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
PREFIX rdfs: http://www.w3.org/2000/01/rdf-schema#
PREFIX drugbank: <http://bio2rdf.org/drugbank:>
PREFIX drugbank_vocabulary: <http://bio2rdf.org/drugbank_vocabulary:>

The linked data network expands
with every reference
drugbank:DB00586
pharmgkb_vocabulary:Drug
rdf:type
rdfs:label
diclofenac [drugbank:DB00586]
pharmgkb:PA449293
drugbank_vocabulary:Drug
pharmgkb_vocabulary:x-drugbank
diclofenac [pharmgkb:PA449293]
rdfs:label
DrugBank
PharmGKB
6

We are building a massive network of linked open data
7
Linking Open Data cloud diagram, by Richard Cyganiak and Anja Jentzsch. http://lod-cloud.net/”

Linked Data for the Life Sciences
• Free and open source
• Leverages Semantic Web standards
• 10B+ interlinked statements from 30+
conventional and high value datasets
• Partnerships with EBI, SIB, NCBI, DBCLS, NCBO,
OpenPHACTS, and many others
chemicals/drugs/formulations,
genomes/genes/proteins, domains
Interactions, complexes & pathways
animal models and phenotypes
Disease, genetic markers, treatments
Terminologies & publications
Alison Callahan, Jose Cruz-Toledo, Peter Ansell, Michel Dumontier:
Bio2RDF Release 2: Improved Coverage, Interoperability and
Provenance of Life Science Linked Data. ESWC 2013: 200-212

Semantic Web
for Health Care and Life Sciences Interest Group (HCLS)
• Mission: to develop, advocate for, and support the use of Semantic
Web technologies across health care, life sciences, clinical research and
translational medicine.
• Since 2001. 86 members from 29 organizations.
• Chairs: Michel Dumontier and Charlie Mead
• Objectives:
– Develop high level and architectural vocabularies.
– Implement proof-of-concept demonstrations and industry-ready
code.
– Document guidelines to accelerate the adoption of the technology.
– Disseminate information about the group's work at government,
industry, academic events and by participating in community
initiatives.

Challenge: Working with Web Data
• Often have inadequate descriptions so we don’t know
what they are about or how they were constructed.
• datasets change over time, but often don’t come with
versioning information
• may have been constructed using other data, but it’s not
clear which version of data was used or whether these
were modified
• Data may be available in a variety of formats
• There may be multiple copies of data from different
providers, but it’s unclear if they are exact copies or
derivatives

Data registries aren’t in sync
– Identifiers.org, Bio2RDF.org, BioSharing.org, etc.
– May be concerned about only some data
elements i.e. incomplete
– May be out-of-date and there is no easy way to
exchange data descriptions
– May contain conflicting information, unclear the
sources used.

no single vocabulary provides all key
metadata fields

Key Use Cases
1. Dataset Identification, Description, Licensing and
Provenance
2. Dataset Discovery (via Catalog)
3. Exchange of Dataset Descriptions
4. Dataset Linking
5. Content Summary
6. Monitoring of Dataset Changes

Objective
• Develop a guidance note for reusing existing
vocabularies to describe datasets with RDF
– Mandatory, recommended, optional descriptors
– Identifiers
– Versioning
– Attribution
– Provenance
– Content summarization
• Recommend vocabulary-linked attributes and
value sets
• Provide reference editor and validation

Dublin Core Metadata Initiative
Widely used
Broadly applicable
– Documents
– Datasets
✗Generic terms
✗Not comprehensive
✗No required properties
@micheldumontier::CEDAR:Jan
15
“Date: A point or period of time
associated with an event in the
lifecycle of the resource.”

DCAT: Data Catalog
 Separates Dataset and Distribution
✗No versioning
✗No prescribed properties

17
@micheldumontier::CEDAR:Jan
VoID: Vocabulary of Interlinked
Datasets
Metadata carried with data
– Directly embedded: void:inDataset
✗No versioning
✗No checklist of requisite fields
✗Only for RDF data

We compiled a list of metadata fields
used across the community
and then surveyed over 20 vocabularies to see if they
provided relevant metadata elements or value sets
To produce a big spreadsheet that maps metadata needs
with existing vocabularies

Dataset
“A collection of data, available for access or
download in one or more formats”
– DCAT

Included Vocabularies

Three Component Metadata Model:
description – version - distribution

Example of Use

61 metadata elements

Metadata element, description, and
example of use

Metadata Specification
constrained property:value pairs
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL
NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and
"OPTIONAL" in this document are to be interpreted as described in
RFC 2119 [RFC2119].

Description
• Identifiers
• Title
• Description
• Homepage
• License
• Language
• Keywords
• Concepts and vocabularies used
• Standards
• Publication

Attribution
• Simple Model
– Individuals are related to roles using specific
properties
e.g. dct:creator, pav:createdBy, pav:curatedBy
• Expandable Model
– Individuals are related to roles and dates by
associated object
– PROV, ViVo

Provenance and Change
• Version number
• Source
• Provenance: retrieved from, derived from,
created with
• Frequency of change

Availability
• Format
• Download URL
• Landing page
• SPARQL endpoint

RDF Dataset Statistics
Basic Statistics
• # of triples
• # of typed entities
• # of distinct subjects
• # of distinct predicates
• # of distinct objects
• # of classes
• # of literals
Enhanced Statistics
• Classes + #
• Properties + triples
• Subject Types + # Property +
triples
• Object Types + # Property +
triples
• Literals + # Property +
triples
• Dataset-Dataset links

Application scenarios

VoID Editor

Validator
New version
using ShEx in
development

Towards Semantic Interoperability

dumontierlab.com
michel.dumontier@stanford.edu
Website: http://dumontierlab.com
Presentations: http://slideshare.com/micheldumontier
37
HCLS:
http://www.w3.org/blog/hcls/
Mailing list:
http://lists.w3.org/Archives/Public/public-semweb-lifesci/
Editors’ Draft:
http://tiny.cc/hcls-datadesc-ed
W3C Interest Group Note:
http://tiny.cc/hcls-datadesc
Special thanks to Alasdair Gray, Scott Marshall, Joachim Baran
Thanks to all other contributors to the HCLS note

W3C HCLS Dataset Description Guidelines

More Related Content

What's hot

Viewers also liked

Similar to W3C HCLS Dataset Description Guidelines

More from Michel Dumontier

Recently uploaded

W3C HCLS Dataset Description Guidelines

Editor's Notes