Clarification specimen role

[cuboideum]

In the project meeting on 19 July 2016, we discussed the use of class obo:OBI_0000112 ('Specimen role') and why it is there in the first place. I have looked at the OBI again to find out more.

Any material entity (obo:BFO_0000040) that bears the specimen role is automatically an instance of the class obo:OBI_0100051 ('Specimen'):

<owl:Class rdf:about="http://purl.obolibrary.org/obo/OBI_0100051"><!-- specimen -->
        <rdfs:label xml:lang="en">specimen</rdfs:label>
        <owl:equivalentClass>
            <owl:Class>
                <owl:intersectionOf rdf:parseType="Collection">
                    <rdf:Description rdf:about="http://purl.obolibrary.org/obo/BFO_0000040"/><!-- material entity -->
                    <owl:Restriction>
                        <owl:onProperty rdf:resource="http://purl.obolibrary.org/obo/RO_0000087"/><!-- has role -->
                        <owl:someValuesFrom rdf:resource="http://purl.obolibrary.org/obo/OBI_0000112"/><!-- specimen role -->
                    </owl:Restriction>
                </owl:intersectionOf>
            </owl:Class>
        </owl:equivalentClass>
</owl:Class>

So there are two ways to model specimen collection:

1. Assign a new instance of obo:OBI_0100051 ('Specimen') for the output of obo:OBI_0000659 ('Specimen collection process').
2. Let the input of obo:OBI_0000659 ('Specimen collection process') bear an instance of obo:OBI_0000112 ('Specimen role') to mark it as the process output.

In terms of RDF reasoning, both solutions are equally valid. In practice, both have their advantages and disadvantages. It remains to assess which solution is more suitable for software integration.

Which approach should we use for RDFBones?

[cuboideum]

The purpose of defining specimen is to distinguish a material entity that has gone through some kind of collection process. Other studies can select their material from among the specimen without repeating the specimen collection process.

From my perspective, these are some arguments for and against the two approaches.

Approach 1: Creating instances of obo:OBI_0100051 ('Specimen')

pros:

• better for modeling collection processes where parts of some material entity are extracted (several specimen from one material entity, e.g. thin sections) or the material entity is modified (e.g. mounted)
• useful in blinded studies

cons:

• many instances are created
• query needed to link specimen to material entity

Approach 2: Assigning specimen role (obo:OBI_0000112)

pros:

• better for modeling collection processes where objects are selected without undergoing any modification
• reduces data, keeps things simple

cons:

• probes are not addressed as separate entities and are difficult to refer to when reused

---

RDFBones should stick to one of the two approaches in order to create absolutely identical data structures across software extensions.

Given the arguments summarised above, and after discussing the issue with @zarquon42b , I do advocate approach 1.

[cuboideum]

This issue was indirectly discussed during the RDFBones/AnthroGraph project meeting on 18 February 2021. According to the background information discussed in issue #121, collection items in a (at least partly) skeletal collection should be specimens derived from skeletal material, not the skeletal material itself.

Therefore, investigations can have the collection specimens as input or they can choose other specimens that are the output of some specimen collection process. These processes have specimens both as input and as output.

For now, the concept of 'specimen role' is not needed with RDFBones.