Assertions/Evidence Model + alignment with Basal Ganglia

[tekrajchhetri]

Moving discussions here so that we could track things and keep everything organized.

Previously agreed:

• Make sio:Annotation a different class as opposed to subclass of sepio:Assertion -- by @tekrajchhetri.

  

  Further to make it more specific and avoid any confusion, we agreed to have the brainkb:DataAnnotation class instead of sio:Annotation as suggested by @patrick-lloyd-ray.

• Update the current eco:Evidence class to use the sepio:EvidenceLine -- suggested by @djarecka.

  Definition sepio:EvidenceLine: An evidence line represents an independent and meaningful argument for or against a particular proposition, that is based on the interpretation of one or more pieces of information as evidence.

Resources:

• Google schema sheet
  https://docs.google.com/spreadsheets/d/1UpjNNJz9o1TXpPS245tbPsVbIx_p5_m2f54KoIwGF8g/edit?gid=1224768617#gid=1224768617

• Definitions
  https://docs.google.com/spreadsheets/d/1R8TShB5XTqBVo0nbrvYjMPtEaXrEVKeqcoHKf8GpIQU/edit?usp=sharing

• Draw.io
  https://drive.google.com/file/d/1oJQDoM-eK3OZqdtD-SWAv7UpCXkP_hVf/view?usp=sharing

Suggestions
@tekrajchhetri It is suggested that we re-use the following ontologies as they align with BFO.

1. ECO
2. SEPIO
3. SIO

Suggested by @satra
4. For provenance use PROV ontology.
5. It would be beneficial for our model to adopt a structure similar to that depicted in Figure 1 of Gundersen, O.E. (2021), The Fundamental Principles of Reproducibility, Philosophical Transactions of the Royal Society A, 379(2197), p.20200210.

To do:

• Update the model (https://brain-bican.github.io/models/index_assertion_evidence/) to include the updated changes.
• Alignment with https://github.com/AllenInstitute/AllenInstituteTaxonomy/tree/main/annotations -- @UCDNJJ

For our discussion, I would propose the following structure.

• Title: Add/Remove/Modify Class/Property – [Name of Class/Property]

• Description: A concise description of the proposal, including whether it pertains to adding, removing, or modifying a class or property in the ontology.

• Proposed Change

  • Name of Class/Property:
    Specify the class or property name. For modifications, mention the current name and proposed name (if it changes).

  • Definition:
    Provide the proposed definition for the class/property. Include current and revised definitions for modifications.

    Examples (where applicable):

    Example 1: [short illustrative example]
    Example 2: [short illustrative example]

    Note: If the proposed changes comes from existing ontologies or vocabularies, it should clearly indicate the name of the ontology and the base IRI.

• Reasoning

  • Why is this change necessary?
    Provide a clear explanation of the motivation behind the change, including gaps in the current ontology or benefits of the change.

    Examples:
    Aligning with best practices.

  • Impact of Change
    Discuss potential impacts on the ontology, such as breaking changes.

[tekrajchhetri]

Proposal to add new classes/property

I propose the following new classes/properties to be added in our assertion/evidence model. This addition will allow us to record evidence as well as align with (or help standardise) Basil ganglia schema https://github.com/AllenInstitute/AllenInstituteTaxonomy/tree/main/annotations].

Proposed Change

Name of Class/Property:

1. add datacite:identifier

   Definition:
   The Identifier is a unique string that identifies a resource. Example: DOI.

2. add datacite:alternateIdentifier

   Definition:
   An identifier other than the primary Identifier applied to the resource being registered. This may be any alphanumeric string which is unique within its domain of issue. May be used for local identifiers, a serial number of an instrument or an inventory number. The AlternateIdentifier should be an additional identifier for the same instance of the resource (i.e., same location, same file).

3. add datacite:format

   Definition:
   Technical format of the resource. Example: Pdf, XML, ...

Alternatives: If we want to stay with the PROV ontology, we could also use the prov:atLocation instead of datacite:identifier

4. add _ ORCID_

   Definition:
   A standard alphanumeric code to uniquely identify scientific and other academic authors and contributors.

   IRI: https://w3id.org/reproduceme#ORCID

   Domain:

   • Prov:Agent

Alternatives: If we want to stay with the PROV ontology, we could also use the prov:wasAttributedTo.

Reasoning

Why is this change necessary?

• It helps to capture the user information and also the details about the evidence, such as the type of the file. Moreover, it will also align with the basal ganglia schema, in particular the following classes:

Basal Ganglia	Mapped Assertion/Evidence Class
powerpoint_url	datacite:identifier (or datacite:alternateIdentifier) + datacite:format
reference	datacite:identifier
expert_ORCID	ORCID
expert_description	brainkb:DataAnnotation

Impact of Change
No breaking changes.

I will also create a figure and update it later.

[tekrajchhetri]

Update:

Following the EDAM - The data analysis and management ontology I would further suggest have some specific classes for identifier. We can already re-use EDAM one including the identifier replacing dcite:identifier. As you can see from the below screenshot, it's very rich and also contains other genomic and biological terms.

[tekrajchhetri]

Figure below shows the updated view of our ontology (not complete but a snippet) with the Identifier class from EDAM