Adding new triage categories for PHI-Canto

[jseager7]

(Requested by @CuzickA)

We have a few publications in PHI-Canto that don't seem to fit into Canto's existing triage categories.

The first case is when a curator has added a publication to PHI-Canto but has not proceeded past the page where they enter their email or name, so we don't know for sure whether the author is the curator of the session.

The second case is when a publication has been added by an author that we have contacted, but again, they've not proceeded past the page where they enter their email or name. We'd like some way to track the publications that we need to follow up with authors.

I was considering adding a triage category along the lines of 'Curatable publications with no author response', but there's a slight technical problem: unless this triage category is added for every version of Canto, I'll presumably have to copy the entire list of triage categories into canto_deploy.yaml, which will put us out of sync with any changes to triage categories in future. Although if they don't change often, then that shouldn't be a big problem.

@ValWood How does PomBase handle cases like this?

[kimrutherford]

Hi James.

The second case is when a publication has been added by an author that we have contacted, but again, they've not proceeded past the page where they enter their email or name. We'd like some way to track the publications that we need to follow up with authors.

In PHI-Canto do you have a report page like this of sessions where the curator hasn't started curating?:
https://curation.pombase.org/pombe/view/list/publications_with_new_sessions?model=track

The "Session sent count" will tell you how often the session has been re-sent to the community curator. If the count is 0 it probably means that a curator entered a PubMed ID in the box on the front page.

Does that help?

Here's what it looks like for us:

There's also a "List of publications with active sessions" which will show sessions that are in the SESSION_ACCEPTED state (session started, no genes added) or CURATION_IN_PROGRESS state:

[CuzickA]

Thanks for the screenshots @kimrutherford

I think you send out curation requests to curators via Canto. We haven't done this yet in PHI-Canto. When we contacted potential community curators we sent them email invites from our Curation email address but not via PHI-Canto. Therefore I don't think we will have any entries in these sections

Perhaps this is something we will need to do in the future with PHI-Canto.

[kimrutherford]

but there's a slight technical problem: unless this triage category is added for every version of Canto, I'll presumably have to copy the entire list of triage categories into canto_deploy.yaml

Hi James. Sorry for the slow reply to this bit.

The triage categories are stored in the database so they can be added at any time with a command like:

./script/canto_add.pl --cvterm "Canto publication triage status" "Curatable publications with no author response" Canto:Curatable_publications_with_no_author_response

[jseager7]

@kimrutherford Thanks for the information. How does adding a triage category through the command line relate to the triage categories stored in canto.yaml?

The canto.yaml file implies this its triage codes are processed at initialisation time (the parent property is named db_initial_data), so I'm guessing this configuration just stores triage categories that will be automatically loaded into Canto at initialisation?

The reason I'm asking is that if the triage configuration is required even for codes that have been added via the command line, then adding triage codes via the command line won't solve the problem I mentioned in my original comment about having to "copy the entire list of triage categories into canto_deploy.yaml".

[kimrutherford]

The canto.yaml file implies this its triage codes are processed at initialisation time

Yep!

The configuration is used when the Canto instance is initialised and ignored from then on.

"copy the entire list of triage categories into canto_deploy.yaml"

That's not necessary.

Everything under "db_initial_data" in the config is only used once, at instance initialisation time.

[jseager7]

@CuzickA Since it's possible to add custom triage categories to PHI-Canto solely for our own use, would you prefer to have dedicated triage categories for the following cases?

• A curator we have not contacted has added a publication to PHI-Canto but has not proceeded past the Curator Details page (in which case we can only guess who the curator is).

• An author we have contacted has added a publication to PHI-Canto but has not proceeded past the Curator Details page (in which case we need to follow this up with the author).

[jseager7]

The PHI-base team have agreed to add dedicated triage categories for these cases, named something like:

• Invited author, no response
• Uninvited author, no response

We were also considering adding reports for these triage categories, named something like:

• Curatable publications with invited author without sessions
• Curatable publications with uninvited author without sessions

@CuzickA Does that look okay to you? Would you rather use 'curator' instead of 'author'?

@kimrutherford Can reports be added from the command line in the same way that triage categories can? If not, I'll probably have to copy the class_info configuration object out of canto.yaml before patching in the extra reports for these triage codes.

At least the configuration itself seems straightforward. I'm guessing it would be like this:

curatable_publications_with_invited_author_without_sessions:
      class_display_name: Curatable publications with invited author without sessions
      extends: pub
      constraint: "triage_status_id in (select cvterm_id from cvterm where name = 'Invited author, no response')"

---

Note that all this manual classification presumably wouldn't be necessary if we used Canto to send out curation invitations in the same way that the PomBase team do, since then the cases would probably fit into the existing reports. Unfortunately, I'm not sure we'll be able to follow PomBase's practices since we're planning on recruiting journals to send curation invitations to authors on behalf of PHI-base as part of their article submission process.

[CuzickA]

Invited author, no response
Uninvited author, no response

instead of 'no response' would it be better to have 'session not started'?

I think that 'curator' may be better than 'author'.

Just to note that in both cases the curator will probably not have entered their name and email address, so we will have to match the PMID entered up against the curator invite emails to know whether to triage into the invited or uninvited category.

[jseager7]

instead of 'no response' would it be better to have 'session not started'?

That's fine with me. So the codes and reports will be named:

• Invited curator, session not started
• Uninvited curator, session not started
• Curatable publications with invited curator without sessions
• Curatable publications with uninvited curator without sessions

The triage labels are a bit long, but I've tested them in the admin interface and there don't seem to be any display issues with the width.

[CuzickA]

Looks good thanks