global rdf descriptors for germplasm data
DESCRIPTION
Presentation delivered in the context of the Agricultural Data Interoperability WG meeeting, during the RDA 3rd Plenary Meeting in Dublin, Ireland. 26/3/2014. The presentation is mostly focused on the work done by the agINFRA project towards proposing a methodology for the definition of Germplasm descriptors as RDF, based on the existing work of experts in the field and making use of the existing effort in this direction.TRANSCRIPT
Global RDF Descriptors for Germplasm Data
Vassilis ProtonotariosAgricultural Biotechnologist, PhD
Agro-Know, Greece
RDA 3° Plenary Meeting, Dublin, IrelandAgricultural Data Interoperability Group Meeting
Background
Connecting the pieces
agINFRA Germplasm Working Group
Agricultural Data Interoperability IG
Germplasm Data Analysis
Agricultural linked data layer
The agINFRA project
• A project funded under the FP7 program of EC• Consortium with expertise on– Technology / infrastructures– Data / data management
Combined to facilitate agricultural data sharingMore info at:
www.aginfra.eu
The agINFRA project
• Aims to enhance the interoperability between the agricultural data sources
– Data sharing by• Metadata aggregation & linking data• Design and deploy the linked ag-data framework
– Methodology for linking data– Provide the infrastructure needed• Both cloud- and grid-based services• Tools, APIs etc.
agINFRA major data types
agINFRA
Bibliographic
Agri Statistics & Economics
Educational
Germplasm
Soil data
Profiles
Raw data
Other?
Focusing on germplasm
Local Databases
National DatabasesAggregators
GENESYSEURISCO
GBIF
Italian
Italian University
Italian research center
Chinese Chinese research center
Data flow
Focusing on germplasm
Local Databases
National DatabasesAggregators
GENESYSEURISCO
Italian
Italian University
Italian research center
Chinese Chinese research center
The issue ?
• Heterogeneity!– Data types– Data formats– Data management workflows– Standards used– Metadata exposure options– ….
• Lack of connectivity with other data sources
The agINFRA Germplasm Working Group
The Germplasm Working Group
• Created in the context of the agINFRA project• Initially included agINFRA stakeholders– now expanded to host all stakeholders
• The group is NOT a group of experts on germplasm data!
The scope of the agINFRA Germplasm WG
• Enable/enhance interoperability between germplasm databases – By developing the services for
• exchanging their data and • delivering their data to other partners
• Focusing on three actions:1. Identify 2. Organize & Reuse3. Propose
agINFRA Germplasm WG objectives
• IDENTIFY: collect all information related to germplasm data– People/groups– Namespaces (metadata, KOS)– Standards– Workflows– Events
• ORGANIZE & REUSE: engage all stakeholders & available resources, analyze existing standards , facilitate collaboration
• PROPOSE: linked data framework to connect data sources
• facilitate data sharing between germplasm data sources
Germplasm related information
data management
workflows
metadata schemas
Working groups in
germplasm
Events (for connecting stakeholders)
KOS (ontologies,
thesauri, vocabularies
etc.)
Data exposure capabilities
Germplasm related information
data management
workflows
metadata schemas
Working groups in
germplasm
Events (for connecting stakeholders)
KOS (ontologies,
thesauri, vocabularies
etc.)
Data exposure capabilities
The Germplasm WG wiki
• Central point of reference
• Freely accessible (no login required)
http://wiki.aginfra.eu/index.php/Germplasm_Working_Group
Key outcomes of the group (1)
Dossier on Germplasm Information:– Major programs– Major information systems and services– agINFRA germplasm data sources (CGRIS & CRA)– Core standards for germplasm information – Plant nomenclature, taxonomies and ontologies– Plant genomic resources– Related references and links
• Freely available from the Germplasm Group wiki
Key outcomes of the group (2)
Key outcomes of the group (3)
• Speakers from key players in the biodiversity data field– GBIF, EURISCO, GENESYS, CGIAR, EGFAR, CRA etc.
• Aimed to provide the basis for the linked germplasm data framework
Existing work
DwC-G KOSs
• Germplasm Term Vocabulary • A vocabulary of terms for describing and annotating
germplasm information resources– http://purl.org/germplasm/germplasmTerm#TERM
• Germplasm Type vocabulary• List of controlled values for some of the germplasm terms
– http://purl.org/germplasm/germplasmType#TYPE
• Germplasm ontology• to digitize and provide persistent identifiers for the terms
contained within the PGR Descriptors publications– http://purl.org/germplasm/ontology
DwC-G linked data
DwC-SW
• An ontology using Darwin Core terms to make it possible to describe biodiversity resources in the Semantic Web.
https://code.google.com/p/darwin-sw
Bioversity Crop Descriptors
• Crop Descriptors– Provide an international format and a universally understood
language for plant genetic resources data. – They are targeted at farmers, curators, breeders, scientists and users
and facilitate the exchange and use of resources. – Information includes such details as the plant's height, flowering
patterns and ancestral history.• FAO/Bioversity Multi-crop Passport Descriptors (MCPD)
– Originally published in 2001– widely used as the international standard to facilitate germplasm
passport information exchange. – Now expanded to include emerging documentation needs, this new
version resulted from consultation with more than 300 scientists from 187 institutions in 87 countries.
Wheat descriptors
• Descriptors for wheat and Aegilops (1978)• Descriptors for wheat (Revised) (1985)
– Not available as Linked Data
Methodology: towards the RDF germplasm descriptors
Linked Data vocabularies
• Metadata vocabularies: Metadata sets, metadata element sets – they provide metadata elements to describe individual pieces of
information in the data sets.– Example: Dublin Core is a vocabulary that prescribes the property
dc:date for the publishing date of a document.
• Value vocabularies (KOS): Controlled vocabularies, authority data– they provide sets of values for (some of) the metadata elements.– Example: AGROVOC provides a set of values for agricultural topics
that can be used as values for the dc:subject property.
LOD guidelines (Berners Lee, 2006)
1.“Use URIs as names for things”– concepts / values in value vocabularies and classes and properties in description vocabularies, as well
as the vocabularies themselves, have to be identified by URIs.
2.“Use HTTP URIs so that people can look up those names”– the URIs for concept / values, classes and properties, as well as vocabularies, have to be resolved as
HTTP URLs.
3.“When someone looks up a URI, provide useful information”– the URLs for concepts, classes and properties, as well as vocabularies, have to return an HTML page
with useful information when requested by browsers, or RDF when requested by RDF software; besides, vocabularies should be available for querying behind a SPARQL endpoint.
4.“Include links to other URIs, so that more things can be discovered”– the URIs of concepts, classes and properties should whenever possible be linked to URIs in other
vocabularies, for instance as close match of another concept or sub-class of another class.
Proposed methodology
1. Analyze metadata schemas & KOSs used to describe germplasm resources
2. Define attributes & vocabularies that can be used to expose germplasm resources in linked data format.
3. Provide a set of recommendations for the exposure of germplasm resources as linked data
4. Embed the recommendations in the data infrastructure of agINFRA – to allow the exposure of germplasm resources as LOD.
The next steps
Application of the linked agricultural data framework in germplasm
1. Definition of base schema– Darwin Core for Germplasm to be used as base
schema• Already available in SKOS• Vocabularies published as linked data
– Germplasm Vocabularies• Germplasm Term Vocabulary • Germplasm Type Vocabulary
– Germplasm ontology
2. Publication of local classifications / lists for germplasm as LOD KOSs – if possible use DwC Types directly
Application of the linked agricultural data framework in germplasm
3. Linking of terms in new KOSs to terms in existing KOSs – e.g. DwC Types, AGROVOC
4. Link CAAS and CRA germplasm records using scientific name > AGROVOC
5. Collaboration with technical partners– technical specifications on how to write procedures that extract the
relevant data from the database and "triplify" them (i.e. both serialize them as RDF and use URIs instead of just strings whenever possible, also linking to AGROVOC URIs when possible).
…and more next steps (optional)
• Update the existing analysis with new data• Collect new user requirements• (re)define the mappings between metadata
schemas and KOSs (if needed)• Fine-tune the linked data approach
Time plan
Time plan
• June 2014: Germplasm vocabularies– Metadata model: Darwin Core SW + DwC-G as the
reference• Publish local classifications / lists for germplasm as LOD
KOSs (if possible use DwC Types directly)• Link terms in new KOSs to terms in existing KOSs (e.g.
DwC Types, AGROVOC)• Germplasm phenotypic values / classifications linked to
Phenotypic Ontology terms?
Time plan
• August 2014: Germplasm RDF– Expose some RDF output and API access for
germplasm datasets (basic DwC RDF, essentially basic passport descriptors).
– Mandatory data for interlinking: scientific name OR AGROVOC term
Time plan
• October 2014: Consuming data from agINFRA services and components– Link CGRIS and CRA germplasm records using
scientific name > AGROVOC
Source: http://verastic.com/social/why-do-people-not-say-thank-you.html