search

geneontology / noctua

Graph-based modeling environment for biology, including prototype editor and services
http://noctua.geneontology.org/
BSD 3-Clause "New" or "Revised" License
37 stars 13 forks source link

Further constrain the ECO subset available #856

Open cmungall opened 8 months ago

cmungall commented 8 months ago

Currently we use an ECO subset that restricts to "manual" terms.

This includes terms that are likely too narrow experimentally:

ECO:0005652 ! methidiumpropyl-ethylenediaminetetraacetic acid iron (II) footprinting evidence used in manual assertion ECO:0006041 ! genetically encoded fluorescent ion concentration sensor assay evidence used in manual assertion

As well as narrow subclasses of IC

ECO:0007756 ! curator inference from Britannica used in manual assertion ECO:0007754 ! curator inference from Wikipedia used in manual assertion ECO:0007755 ! curator inference from MedlinePlus encyclopedia used in manual assertion

In theory it is fine is people want to use granular ECO terms, these will all roll up the GAF codes

But is this actually our policy? If we allow people to use any granularity then the majority will annotate to broad terms that are equivalent to GAF codes and the end results will be frustrating to use for users that do want granular codes, because the annotation will be at inconsistent levels.

We should make an intentional decision about whether which levels of granularity are MUST, SHOULD, SHOULD NOT, MAY, and reflect this in the subsets available to editors.

vanaukenk commented 7 months ago

We asked curators about ECO usage on the 2024-01-30 annotation call. Some groups do use some of the more granular codes. For example: Xenbase - more granular IMP codes CACAO - easier for students to use the more granular, methods-specific evidence

Let's prioritize this issue and associated ECO work on the 2024-01-14 managers/projects call.