"Interacting species" should be captured as an extension with the "has input" relation

[pgaudet]

Hello,

For annotations to children of GO:0044419 interspecies interaction between organisms, we now recommend to start migrating away from using the 'interacting taxon' field of GAF/GPAD, and instead use an annotation extension with the "has input" relation.

This will be further discussed on annotation calls.

Thanks, Pascale

[hattrill]

@alexsign adding you to let you know of changes

[alexsign]

@hattrill thanks! I'll tag @pgarmiri as well

[pgarmiri]

Hi,

The suggested change, technically, is possible and biologically makes sense. My only concern about this is that not all groups use annotation extensions and they are not recognized by many tools. Many annotations are going to be missed. Both because people can't add AE (so no new annotations) and old annotations will not be retrievable. A point for the discussion.

Thanks,

Penelope

[pgaudet]

Are they using the interacting taxon field ? It would be awesome to know what people do use !! Please let us know if you have any data about that.

Thanks @pgarmiri

[ValWood]

I didn't know about this! I'm just getting the dual taxon part implemented in Canto. It seems that we might need to do something different?

[pgarmiri]

Hi @pgaudet , Just got some numbers from Alex. So, here is the number of annotations with interacting taxon for each group.

UniProt 1749 WB 157 CAFA 67 ParkinsonsUK-UCL 28 ARUK-UCL 23 BHF-UCL 21 DIBU 3 GO_Central 1 MTBBASE 1

Thanks,

Penelope

[vanaukenk]

Thank you @pgarmiri We'll discuss on today's annotation call

[pgarmiri]

Hi @pgaudet ,

I realized that some important numbers were missing. The breakdown of the UniProt annotations. So, here is the number of annotations with interacting taxon for each of the UniProt group.

Total UniProt 1749 UniProt SWIE 1687 UniProt SWIS 60 UniProt SWIP 2

At SIB they don't use annotation extensions, so as the things are, they would not be able to capture this information. Technically, this is something that can change but I am not in position to talk about this.

What is your take on this @sylvainpoux ?

Could you please also tag Ceci at PIR? They are at the same situation.

Thanks,

Penelope

[pgaudet]

It would be nice if SWIS would start to use extensions. Now that the documentation is better maybe we can consider it ?

[sylvainpoux]

Hi Pascale,

sure. The reason why we never used extensions is because we have never been trained. @pgaudet, could you organize the training for curators and check our annotations?

For existing SWIS annotations, the best would be to contact curators who made them and explain them what to do

Thanks

Sylvain

[pgaudet]

Yes @sylvainpoux , I'd be happy to organize a training session.

[suzialeksander]

Are we still going this direction? Neither GPAD 2.0 nor GAF 2.2 seem to indicate we're discouraging use of the taxon field. Either way there's nothing actionable here, if this is just an announcement it's OOD.