#6 is_a relationships to be added as well as cross-products

closed
nobody
None
5
2011-02-17
2011-02-01
Emily Dimmer
No

In the current draft of ECO there are cross-products between terms under 'assertion method' and 'evidence'. Is it envisaged that all terms under assertion method will eventually have an is_a relationship to one of the terms in the evidence ontology?

Thanks,
Emily

Discussion

  • There are no terms that are under assertion method (except automatic assertion and manual assertion themselves) that do not have is_a relationship to an evidence term. Here are some further details:

    The current release of ECO has two root classes, "evidence" and "assertion method". Assertion method has two child terms, "manual assertion" and "automatic assertion", the latter of which was formerly "inferred from electronic annotation" (IEA). (IEA as a term obscured the knowledge of what actual analysis was performed, as "electronic" is not particularly meaningful.)

    In order to capture the idea that an annotation was not reviewed by an annotator, one would use a term that combines evidence type and automatic assertion. For example, an unreviewed sequence similarity analysis would be captured with "sequence similarity used in automatic assertion" (ECO:0000249). In keeping with the original structure of the ontology, this term has an is_a relationship with "sequence similarity". However, its relationship to "automatic assertion" is used_in, because it is not a type of automatic assertion, per se, but rather a type of evidence that is used in making an assertion.

    If you were to open ECO in OBO-Edit, you would see the explicit is_a relationships for evidence types. If you were to turn on the reasoner, you would see that the used_in relationships to assertion method are inferred, because the combinatorial (evidence x assertion method) terms are cross products. In this example, ECO:0000249 has the intersection genus "sequence similarity" and discriminating relation "used_in" "automatic assertion".

    Ultimately, all terms under the evidence root will have a used_in relationship to "manual assertion" under the assertion method root. However, only some evidence types will have the used_in relationship to "automatic assertion". For example, "inference from background scientific knowledge" (synonym "inferred by curator") would not be used_in automatic assertion, as this is always a manual process. Thus, the term "inference from background scientific knowledge used in manual assertion" exists, but "inference from background scientific knowledge used in automatic assertion" does not.

     
  • Emily Dimmer
    Emily Dimmer
    2011-02-17

    • status: open --> closed
     
  • Emily Dimmer
    Emily Dimmer
    2011-02-17

    I see, thanks for guiding me to the reasoner!
    Cheers,
    Emily