If i want to publish experimental results, preferably via RDFa using a standardized ontology what would be a good source to use. I am thinking of a triple such as:
Protein X --- Interacts with --- Protein Y
Where the ontology would spell out "Interacts with". Are there existing projects that mine the literature and generate triples like this?
•
written
by
Dave Bridges
I would recommend formatting your data using the IMEx (International Molecular Exchange Consortium) curation guidelines. This will allow you to submit your data easily to any of the participant databases (DIP, MINT, INTACT, etc). IMEx uses The PSI (Proteomics Standards Initiative) Molecular Interactions controlled vocabulary. There is a PSI-MI XML/CV validator here.
Yes, there are effort to mine protein-protein interactions and experiemntal methods from the literature, see the BioCreative PPI tasks. In particular, see the Biocreative 3 IMT (Interaction Method Task) which uses the PSI-MI.
this sounds great for protein interactions, but my question was meant more generally.
To add to the answers here - if you want to broaden the understanding of "publishing experimental results" somewhat to "describing an experiment and the results of the experiment" then the Ontology for Biomedical Investigations is also worth considering. It has notions of experiments, planned outcomes of experiments, real outcomes etc. As an example, here are some axioms defining the term "assay" - which come close to experiment:
has_specified_input some
(material_entity
and (has_role some 'evaluant role'))
has_specified_output only 'information content entity'
has_specified_output some
('information content entity'
and ('is about' some
(continuant
and (has_role some 'evaluant role'))))
realizes some 'evaluant role'
realizes some (is_concretization_of some
('plan specification'
and (has_part some 'objective specification')))
There's almost everything there to define an experiment and experimental findings.
•
written
by
Nico Adams •
I agree, a broader view of different experiments is worth considering. Another more general experimental ontology is EXPO: http://expo.sourceforge.net/
for an ontology describing the interaction, look at the BioPortal and search for a vocabulary. For example:
There's plenty of work by people maintaining the databases (see Casey Bergman's answer). If you want to start from scratch and not overly rely on complex schemes, do read the principal considerations by Lu et al. from Mark Gerstein's group, which suggests a plain and easy to implement ontology.
The Experimental Factor Ontology (EFO) is an application focused ontology modelling the experimental factors in ArrayExpress. The ontology has been developed to increase the richness of the annotations that are currently made in the ArrayExpress repository, to promote consistent annotation, to facilitate automatic annotation, and to integrate external data.
The methodology employed in the development of EFO involves construction of mappings to multiple existing domain specific ontologies, such as Disease Ontology and Cell Type Ontology. This is achieved using a combination of automated and manual curation steps.
Log in to answer this question.
+1 for using RDF
thanks to everyone for the great answers. As i was thinking about it some more, it would be nice for all experimental data to somehow be reproducibly put into the web of data. But i think for that one would require not just simple triples, but also some way to point directly to the result. I guess what I am getting at would be some way that you could represent claims/facts in a triple and then be able to add supporting/dissenting experiments to that claim with a second triple.