GO Managers, Weds. Jan. 30, 2008 8a.m. PST, 9a.m. MST, 10a.m. CST, 11a.m. EST, 4p.m. GMT
Present: Midori Harris, Pascale Gaudet, Jennifer Deegan, Chris Mungall, Judy Blake, Michael Ashburner, Suzanna Lewis.
Review Action Items from previous meeting
All: review document on regulation wiki page.
DONE - The mail about the new regulation relationships has now gone to the go list.
There was a discussion of whether people are reacting to the mail. Midori has had some questions at group leaders meeting. David feels we will hear more after implementation. Chris is trying to encourage people in different sites who are part of the software group to think about it.
David gave a talk about regulation at MGI and Tanya gave the same talk at TAIR. It will also given at the consortium meeting.
Action Item: David will send regulation relationship slides round.
Chris: Draft proposal for GO slim maintenance.
There have been discussions about whether we can split the GO slims out of the GO ontology file. However, before we can do this we need changes to the way the GO file is maintained.
We need to have a real editor's version that the users don't have access to. We did try to implement this in gene_ontology_edit.obo (OBO1.2 format) but as that is now available to all the users (as a source of added OBO 1.2 format features) it has backfired and is not a true editor's version.
A summary of Chris's proposal is shown here:
We discussed the options and decided on the following:
- The current gene_ontology.obo and gene_ontology_edit.obo files will be deprecated. They will still be available for people who download them by script for their tools, but they will not be advertised.
- We will widely advertise the existence of two files called gene_ontology.1_0.obo and gene_ontology.1_2.obo:
- gene_ontology.1_0.obo will be identical to gene_ontology.obo except in file name. It will be in OBO 1.0 format.
- gene_ontology.1_2.obo will be identical to gene_ontology_edit.obo except in file name. It will be in OBO 1.2 format.
- We will have a third file that is for the use of editors only and it will be called gene_ontology_unstable.obo. It will not be advertised to users. When we are developing new systems, like starting adding links from molecular function to biological process, we will be able to start adding the links to this file, and they will be stripped out during the conversion to the user's files. This will allow development to happen much more quickly without disrupting the work of the users or necessitating announcements and waiting periods.
Action Item: Software group should implement this and pick a date and time for the editors to switch over to using the new file.
Pascale, Jim et al.: Continue discussing wiki-based annotation tool.
Pascale (and rest of annotators): Continue improving annotation SOPs.
Progress has been made on quality control documentation.
Judy: Contact sea urchin db
- about doing GO annotation: Mike had contact with the Sea Urchin group and they are not ready to annotate yet. They are still assembling the genome. We should contact them again maybe in April or March.
Judy has contacted the group and they said that they are still working on their genome and that we should write to them again in March or April
Action Item: Judy will contact them.
See if we can collaborate with journals to obtain annotations from users.
Responsible groups: managers, software group and annotators?: Test various methods of accepting user submissions for annotation; converge on best solution, and seek journal buy-in.
A report will be given on this in the Outreach section.
Tanya is ready to demo the annotation tool that will be used by Plant Physiology.
Action: ask Tanya if she can demo at the manager's call in a fortnight. Wednesday 13th February 8a.m. PST
Chris: Ask the wider manager group whether the ontology development wiki should be public.
Maybe also there are better ways to link the two wikis?
Consensus: We were all in favour of having just one wiki, with 95% of the pages being publicly viewable.
Action Item: Mike to look into how this can be done.
The regulation relationships are ready to go live.
David Hill, Chris Mungall, and Tanya Berardini made about 1200 changes.
[David and Chris: you explained something about the files at this point but I didn't get it. Could you possibly add it in? It was to do with how one file is denser with positive and negative regulation relationships and the other is topologically identical to the normal GO.]
MGI is test loading the file.
Some people think graph is too tangled with all three relationships and we don't want it to be unworkable so are not sure which way we are going to go. (Not sure whether to use the positive and negative relationships as well as the plain regulates relationship.)
The reasoner is able to find errors now that it couldn't before. Tanya and David are going through the reports it produces and making fixes. Sometimes there are child terms that we don't really want to enumerate.
Also sometimes problems in univocity between process and regulation of process terms. We are sending the reports on this to Larry Hunter.
A lot of the terms are from the blood pressure and muscle terms. This work should take a couple of weeks. Chris has also made other quality control reports.
One thing shown is that regulates relationships don't parallel the plain process structure. They are going to fix that bit. The work also shows that the reasoner really helps the ontology to be better. This is helping a lot in the long run.
The current plan is to go live with the file at end of March. The documentation is being made just now, and will got to Amelia Ireland to go on the web. There will be training courses, and documentation of rules on how to make new terms. This will be the first set of cross-products to go live.
Should we do the cell type cross products or the molecular function (MF) to biological process (BP) cross products next?
There are different sets of people involved in the MF-BP cross products from those involved in the cell type cross products so there is no need to do one and then the other. Also once we have the editor version of the ontology file the systems any file format changes can be implemented incrementally without having to announce them until they are ready to go. In addition the actual cross-products will be in a separate file that is only needed by the editors.
The cell type cross product file is nearly all done and is in the scratch directory. Tanya and David could look through it and get it ready to go live.
Slight concern at lack of manpower to make the needed changes in the cell type ontology, and at the fact that it is external so we have less control. However, it is hoped that if we press ahead then
this might encourage action on cell type terms.
Action: David can start cell type with Tanya when they finishes regulates.
MF-BP cross products
We are not sure about all the tools changes that will be needed for this, however, we could go ahead and start a pilot project and make some links.
It will be best to just start working on the links as until we start actually trying to make links it won't be apparent what software changes are needed.
It is suggested that Harold might start on just a few pathways and try making links to see how it goes, as he is a biochemist. He could work with Reactome on this.
This would be with the assistance of Chris and David for various technical and content issues.
Action: David will ask Harold if he is willing to go ahead and start this. If it works out he could tell us how it went in a month on the managers call.
This text was added as a written report as there was not time to explain it during the meeting:
Jennifer Deegan is preparing a 2 hour tutorial and talk for an Arabidopsis institute in Gent in Belgium for 25th-27th February. She hopes to demo the new GO slimmer tool in AmiGO for 20 people in a hands on workshop.
TAIR at PAG
Donghui Li reports that TAIR organised a workshop on community annotation at the Plant and Animal Genome XVI Conference on 13 January 2008, San Diego, CA.
Representatives from three MODs talked about the current status and future plan in regard to community annotation:
TAIR, SGN (SOL Genomics Network) and WormBase.
TAIR and journal community annotation plans
Tanya Berardini has been in discussion with the Plant Journal to see if they would be interested to ask for GO annotations from users when they submit manuscripts. Kate Dreher at TAIR initiated the
conversation. Also Plant Physiology, who are already working with TAIR to accept annotations from authors, is very close to starting accepting community annotations. They also have a draft of the editorial that they will publish to introduce the system and the tool is ready in it's first phase, and now being improved. .
GOA and Transcriptomics Meeting
Rachael Huntley has been teaching at an EBI training course 'Transcriptomics' 28-30th Jan 2008.
TAIR, Cornell and Sol Genomics Network database
Groups at TAIR and Cornell (including staff trained at Gramene) are working with the Sol Genomics Network database (http://www.sgn.cornell.edu) to convert their manual annotation for submission to the GO consortium. The staff at SOL involved are Dr Naama Menda and Dr Anuradha Pujar.
Nature Connotea tool
Jennifer Deegan attended a talk at EBI on the new Nature Connotea tool. This tool allows scientists to tag publications of interest and group them by topic and label them with text descriptions. Connotea is open source and plugins can easily be written. There is already a plugin for firefox to allow users to label publications with terms from the GO ontologies. Jennifer enquired about whether this might be extended to allow users to label specific parts of the text with GO terms, or in some other way to support community annotations. The speaker (Ian Mulvany) was interested in the idea and said that they would add this to their very long feature request list. However he said that if anyone else wanted to write plugins to for this functionality then Nature would be happy to incorporate the plugins into the main programme to open up the features to all users.
Chris Mungall: There is a new dictybase programmer called Siddhartha.
Chris has met him by teleconference.
Previously Sohel had done prototype of the reference genome interface in ruby on rails and chado. The new programmer has had a look at tool interface and Seth has had a look at code. Seth is swamped by AmiGO currently.
There needs to be a steering group to help Siddhartha. They hope to have a rewritten prototype to show Pascale, and others, and to get it looking better to go to reference genomes group. David asks to be in the group with Pascale that looks at the tool.
Meeting on natural language processing mentioned.
Judy would like to start a list of 2008 papers ready for next progress report.
If everyone could keep adding to the list of talks, tutorials, publications etc. throughout the year it would be very much appreciated:
- GOC meeting: April 22-23, Salt Lake City
- GO reference genome meeting: April 20-21, Salt Lake City
- Avian annotation workshop (Fiona): May 19-20, 2008
- Jen gives a tutorial in Belgium (funded by EBI) Feb
- Midori will give a talk at the London School of economics in march
Discussion of Erika Feltrin's muscle paper. Jennifer Deegan and Erika Feltrin wrote to a bioinformatics journal to ask if the paper was of interest to them but got no reply. After discussion there was a recommendation by all to send this to a muscle journal rather than a bioinformatics journal. Send with a covering letter to explain why this is of interest to the users.
Action Item: David will send regulation relationship slides round.
Action Item: Software group should implement new editor's version of ontology file (and other details of the plan) and pick a date and time for the editors to switch over to using the new file.
Action Item: Judy will contact sea urchin db in March or April.
Action Item: ask Tanya if she can demo at the manager's call in a fortnight. Wednesday 13th February 8a.m. PST
Action Item: Mike to look into how GO wikis can be combined.
Action Item : David can start cell type with Tanya when they finishes regulates.
Action Item : David will ask Harold if he is willing to go ahead and start this. If it works out he could tell us how it went in a month on the managers call.