MetWare Aiding metabolite identification (and more) Egon Willighagen et al. Wageningen University – Plant Research International / Biometris 2008-05-08, AB@PRI
Jan 27, 2015
MetWareAiding metabolite identification (and more)
Egon Willighagen et al.
Wageningen University – Plant Research International / Biometris
2008-05-08, AB@PRI
Who are we?
Applied Bioinformatics Group Plant Research International
I Jan van Haarst (JSF pages)
I Joost de Groot (ProtoRAW)
I Mark Fiers (BioMoby services, MetWare integration)
I Velitchka Mihaleva (RT/RI prediction modeling)
I Egon Willighagen (SKOS, integration of data preprocessing)
I Roeland van Ham (group leader)
http://www.ab.wur.nl/
2008-05-08, AB@PRI
The Setting
I Metabolite identityI accurate mass, retention timeI isotope patternI full mass spectrum
I Systems BiologyI if we change this, what happens to the metabolome?
I Where, this :=I genotype, mutationsI transcriptionI growing conditionsI geographical locationI ...
2008-05-08, AB@PRI
The Problems
I BiochemicalI biological variationI detection limitsI rough chemical characterization
I Bio- and ChemoinformaticsI No uniform accessI Many manual steps involvedI Data sizesI Diversity of standards and formats
Information loss!
2008-05-08, AB@PRI
The Solutions
I BiochemicalI wait for next generation devices (not)I know your errorsI maximize information extraction
I Bio- and ChemoinformaticsI improve data extractionI standardized programming interfacesI standardized ontologiesI standardized data exchange formatsI (did you spot the pattern?)
2008-05-08, AB@PRI
Our Solution: MetWare + MetFlow
2008-05-08, AB@PRI
MetWare: the building blocks
The Design
I backed up with an ontology
I conform standards like ArMet
Interfaces ...
I ... towards humans: website (JSF-based)
I ... towards databases: SQL
I ... towards softare: Java Beans
I ... towards workflows: BioMart + webservices
I ... towards statistics software: BioMart
I ... towards other databases: Life Science ID
2008-05-08, AB@PRI
MetWare Design: Modules
I MetRAWI storage facility for RAW metabolomics data
I MetMETAI metadata about DoE, genotypes, illness, conditions, ...
I MetOBSERVI observed metabolites, characterized by RT, mass spectrum,
NMR, ...
I MetIDI know metabolites
I MetCHARI (putative) identity assignments
http://apbioinf001.wurnet.nl/ABWiki/MetWare
http://metware.wiki.sourceforge.net/
2008-05-08, AB@PRI
MetWare Design: Graphical...
2008-05-08, AB@PRI
MetMETA: annotating the wet side
2008-05-08, AB@PRI
MetOBSERV: annotating the dry side
2008-05-08, AB@PRI
MetWare Ontology
I Extends the ArMet Ontology
<skos:Concept rdf:about="Email">
<skos:prefLabel>e-mail</skos:prefLabel>
<skos:definition>The email address of the user.</skos:definition>
<skos:broader rdf:resource="People"/>
<skos:broader rdf:resource="&armet;Email"/>
</skos:Concept>
(No, it does not have to be XML; N3 is fine too. It’s RDF, that’simportant.)
2008-05-08, AB@PRI
Resource Description Framework
I triple
I the subject is a resource
I the object might be a resourceI depending on that, two types of predicates
I resource predicate resourceI resource predicate literal
See also wp:Resource Description Framework
2008-05-08, AB@PRI
Simple Knowledge Organisation System
I XML for ontology definition
I based on RDF and the Web Ontology Language (OWL)I unlike OWL it has loose relations (no is a)
I skos:broaderI skos:narrowerI skos:related
I has a few basic resource-predicate-literal relationsI skos:prefLabelI skos:definition
See also wp:SKOS
2008-05-08, AB@PRI
SPARQL Protocol and RDF Query Language
See also wp:SPARQL
2008-05-08, AB@PRI
Autogeneration of a Java API and unit tests
Why?
I hides the SKOS/XML
I hides the SQL
Advantages of autogeneration
I always synchronized to the ontology and the SQLI automatically generation of unit tests
I who said our group didn’t do unit tests ?!?!
Where are these Java beans used?
I the web GUI: JavaServer Faces
I ...
2008-05-08, AB@PRI
JavaServer Faces: split layout from data model
2008-05-08, AB@PRI
JavaServer Faces: the source
2008-05-08, AB@PRI
MetWare Interface 2: BioMart
I Integration: Taverna, RI http://www.biomart.org/, EBI/CSHL
2008-05-08, AB@PRI
ProtoRAW: the RAW data storage
Joost can tell you all about this...
2008-05-08, AB@PRI
Collaboration
MetWare developmentI Steffen Neumann, Leibniz-Institut fur Pflanzenbiochemie,
Halle, GermanyI our MetWare co-founder in September 2007
http://www.ipb-halle.de/I BioAssist/NMC, Utrecht/Leiden and Mark/Wageningen
I joined in April 2008I will focus on clinical dataI spectral treesI integration with workflow environments
http://www.metabolomicscentre.nl/
2008-05-08, AB@PRI
Conclusions...
I Provides the tools to automate data analysis
I Provides storage for your metabolomics data
I Decreases information loss
http://metware.sf.net/
2008-05-08, AB@PRI