search

oborel / obo-relations

RO is an ontology of relations for use with biological ontologies
http://oborel.github.io/
Other
91 stars 46 forks source link

New term request: ‘positively correlated with’ and ‘negatively correlated with’ #710

Closed CDowland closed 1 year ago

CDowland commented 1 year ago

This request is for two subtypes of the preexisting RO term 'correlated with' (http://purl.obolibrary.org/obo/RO_0002610), which is annotated as follows:

Definition: A relationship that holds between two entities, where the entities exhibit a statistical dependence relationship. The entities may be statistical variables, or they may be other kinds of entities such as diseases, chemical entities or processes.

Curator note: Groups both positive and negative correlation

I’m requesting 'positively correlated with' and 'negatively correlated with' as subtypes of 'correlated with'.

NTR 1:

Preferred term label

positively correlated with

Textual definition

Correlation for which the statistical dependence is positive.

Suggested parent term

RO: correlated with (http://purl.obolibrary.org/obo/RO_0002610)

Attribution

ORCID ID: 0000-0003-1909-9269

NTR 2:

Preferred term label

negatively correlated with

Textual definition

Correlation for which the statistical dependence is negative.

Suggested parent term

RO: correlated with (http://purl.obolibrary.org/obo/RO_0002610)

Attribution

ORCID ID: 0000-0003-1909-9269

As the definition for 'correlated with' notes, a variety of ontological types can occupy the domain and range positions. No domain and range are specified for it, which seems like the right approach for these subtypes as well.

The definitions are based on the definitions of 'correlated with,' adapted to indicate whether the correlation is positive or negative. But perhaps there are clearer/better ways to define them.

wdduncan commented 1 year ago

The definition references the use of statistics, but the label does not. Not all "correlations" are statistical in nature. If you intend this relation to be about things calculated using statistical methods, then we suggest modifying the label to reflect this.

CDowland commented 1 year ago

I'm proposing these as subtypes of RO: 'correlated with', which is indeed a statistical relation, and likewise mentions statistics in the definition but not in the label.

wdduncan commented 1 year ago

@CDowland That is a fair point. However, many on the call were in favor of making a change to the label. Many in the RO community want labels to be more precise moving forward.

A minor suggestion about the textual definition. Modify: 

Correlation for which the statistical dependence is positive.

To something like:

A relation that holds between entities x and y in which an increase in x is correlated with an increase in y.

Also, it would be quite helpful if you (or another member of the OBIB) would join the RO calls when new terms from the OBIB group are requested. It helps move the term approval process more quickly.

CDowland commented 1 year ago

Below is a revised request for the two terms that I hope avoids the concerns raised.

Label: positively correlated with Alternative term: directly correlated with Definition: A relation between entities in which one increases or decreases as the other does the same. Definition source: Suggested parent term: RO: correlated with Characteristics: Symmetric, Reflexive Attribution: ORCID ID: 0000-0003-1909-9269

Label: negatively correlated with Alternative term: inversely correlated with Definition: A relation between entities in which one increases as the other decreases. Definition source: Suggested parent term: RO: correlated with Characteristics: Symmetric, Irreflexive Attribution: ORCID ID: 0000-0003-1909-9269

Also, it would be quite helpful if you (or another member of the OBIB) would join the RO calls when new terms from the OBIB group are requested. It helps move the term approval process more quickly.

This request is not from the OBIB group, but point taken. I’ve now joined the Google group and have the meeting series on my calendar.

wdduncan commented 1 year ago

@bpeters42 and @nataled are you satisfied with the definition? (I am.)

Re 'statistical' in label: The revised definitions don't mention statistics. So, I think the label is fine.

nataled commented 1 year ago

Looks good to me.

bpeters42 commented 1 year ago

Good to me as well.

wdduncan commented 1 year ago

@CDowland You term is approved :)
If I assign an IRI for the term are you (or someone in your group) able to make a PR for this?

CDowland commented 1 year ago

Great, thanks! And yes on the pull request.

wdduncan commented 1 year ago

Thanks @CDowland
Please use id RO_0017003.

CDowland commented 1 year ago

@wdduncan We'll need one more since there are two terms. Thanks.

wdduncan commented 1 year ago

Hah! Good catch @CDowland !
Please also use RO_0017004

wdduncan commented 1 year ago

Close via https://github.com/oborel/obo-relations/pull/719.