EPIC: Data schema discussion ticket

[dosumis]

Proposal for core model:

Extended version of the VFB model for images could work:

(C:Class:Neuron)<-[:INSTANCEOF]-(I:Individual:Neuron)<-[d:depicts]-(J:Image)

We could broaden the image slot to include other data types. Evidence goes on the depicts edge - this could be an individual that aggregates information. Experimental details can be connected to J. I’m hoping we can link to external sources for this, as I worry we don’t have the resources to curate + I’m reluctant to wade in to details of experimental modelling.

With this approach we can add additional axiomatisation to ‘I’ that is useful for query/inference. As we learn more about I, we can add a more precise class.

This needs to be combined with a different approach for definitional (reference) data.

• We already have individuals for dendrogram clusters with an exemplar (definitional) relationship to classes.

@cmungall Comments please: Should be be looking at BioLink instead?

[dosumis]

Update - this will only be relevant if Allen adopts our pipeline infrastructure.