6 January 2015 PAINT Conference Call
January 6th, 2015 PAINT call minutes
Participants: Rama, Chris, Suzi, Pascale, Paul T., Karen, Li, Huaiyu
Test scripts for updates of PAINT GAF files (Suzi)
- Suzi: A script will be finished soon to update the paint GAF files when there are updates on trees (nodes) and literature annotations, etc. It would be helpful if it can be tested by the paint curators. Each one can take a look at a few families in details.
- Huaiyu: This is great. It would be nice if there is a documentation about the script, so that the curators can look at those specifically.
- Suzi: There is a log file that records the updates, such as new nodes or new terms, etc. Maybe that can be used to find the families to look at.
- Huaiyu: PANTHER v.10 should be ready soon. Before it is released, it will be a good test for the script also.
USC FTP SITE IS NOT PROVIDING THE SAME DATA AS THE USC PAINT WEB SERVICE (follow up)
This was discussed at the last call. Anushya generated the modified attr file and was forwarded to Suzi to see if the format is correct. If so, we can modify and regenerate the library.
- Huaiyu sent a sample attr to Suzi.
- Suzi will look at the file, and confirm the format is correct.
- PANTHER group will generate a downloadable library with the modified attr files, and it should include all the other files used by paint, including tree files, msa files and weights files. This library will be put at the permanent location for PAINT software to access.
NOT annotation in PAINT (follow up)
This was discussed at the last call, but not agreement was reached. It was agreed that offline discussion will continue. Just to follow up to see if there is any resolution.
- Paul T. suggested to remove the redundant constraint to allow curators to block a specific function while allowing a more general function.
- Karen used the pol2 example to illustrate the problem.
- Suzi thinks there maybe better way to curate (propagate) instead of using "NOT" in the situation.
- It is agreed that more off-line discussions should be continued. This may require face-to-face meetings.
Are taxon constraints functioning properly in PAINT (Karen)
(This was postponed from the last call.)
- David Hill brought this to my attention. The mouse Capzb gene is getting an IBA annotation to the term 'actin cortical patch' via PAINT, despite the fact that David says that "According to the ontology 'actin cortical patch' is restricted to fungi."
- It's possible the the taxon server was down then - another reason to have touchup running regularly
- To test whether the taxon server is running paste this URL (or substitute your own taxid & GOID as needed)
The Panther family is PTHR10619. It was committed a month ago on Nov 5th.
- The problem could be due to either the taxon server was down or an old version of taxon constraint was used.
- Rama pointed out that this actually happened quite often.
- Karen suggested that a warning should be given when the server was down.
- Suzi suggested to file a feature request to GitHub repository.
- Li asked if there was any Jensins report when something is wrong, such as the server was down.
- Suzi responded that there should be two reports, and Jenkins does reports problems if the server is down.
- Rama: Obviously this is not the case.
- Suzi: will as Heiko to take a look.
(These were postponed from the last call) NOT & IRD (Pascale) NOT Are not allowed when there is a positive annotation in a clade (as far as I can tell). Can this be changed, at least for use with the IRD code ?
- There could be different reasons that a conflicting annotations are going to be made using IRD code. It is better to be captured if such annotations are allowed in PAINT.
- A list of such reasons should be created, and then the feature should be requested in GitHub.
Show PAINTed parent terms in the matrix (Huaiyu) By default, only the most specific terms are listed in the matrix. When a parent term is propagated, it is now visible in the matrix, but the term is listed in the Annotation window. This is a bit confusing. My suggestion is that all PAINTed terms should be shown in the matrix window.
- Suzi agrees that this is a problem that confuses people. Due to all the other requests, this is relatively low on the priority list.