Calling in: Michael Ashburner, Judy Blake, Rex Chisholm, Jennifer Clark, Jane Lomax, Midori Harris, David Hill, Suzi Lewis, Chris Mungall, Mike Cherry
Action Items from August 16th.
(Remaining from July 19) Will we use RCA for InterPro2go, TMHMM, etc? need input from Michelle Gwinn
- no change
Action item: Rex (with help as needed) to put [reference genome gene list] spreadsheet on wiki as table, with new columns as requested for UniProt, HGNC, etc. Run draft by GO managers; notify Reference Genomes list when it's ready.
Action item: Midori to put wish list [for tracker to replace SourceForge] on wiki
Action item: advisory board meeting presenters to prepare talk drafts in time for Aug 30 call
- several done - comments on slides later in meeting.
Action item: Rex to make [disease ontology] file available to GO managers by Aug. 23rd
- Done (CVS info sent)
(Note: last call minutes are on the wiki at http://wiki.geneontology.org/index.php/Managers_16Aug06)
Progress Reports (order arbitrary)
The PAMGO proposal is out for review in SourceForge.
We are working on the is_a complete paths David has circulated a spreadsheet with the names of new terms with no is_a parent. The other columns are to be filled in with the names of the is_a parents. David has done his set of terms, then Cindy did hers, and Doug is looking now.
This is unlikely to be a difficult issue and we hope it will soon be complete. There will be a small number of new terms made and Jen will make all the edits.
If intermediate parents are needed this will be done later. The graph will still be correct without them.
We are planning to work on is_a completeness in the process ontology in Seattle and again in Cambridge when David comes to visit.
We will do this from the top down because the other way has been tried and did not work.
David mentions that is_a completeness in the component ontology is very helpful. He says that currently the cns terms seem very organism specific but once the graph is is_a complete the terms will become applicable across the species.
The immunology graph has taken a long time. This is because it was a huge piece of work and intersected a lot with the PAMGO graph. Getting all the intersections straightened out has taken a long time and been very complicated.
We discussed the difference between the CNS content development and the immunology content development to try to work out lessons for next time.
The CNS work went better because there was a more manageable chunk of biology chosen and it was broken down into sections. So rather than trying to do the whole nervous system we just did some specific parts. The fact that is didn't intersect with any other difficult areas helped a lot. Also the people making the terms needed them for annotation so it was very quick to get feedback. Getting the terms in in the first place was quick because we just did the part_of relationships first and the is_as will go in while the terms are live.
The people working on the immunology section are working hard to get it completed and hope to have it out soon.
David mentions that the PAMGO people made their graph differently from the CNS people. The PAMGO people make the complete is_a path first and the part_of relationships are yet to be filled in. With CNS is was the opposite. They made all the part_of relationships and the is_a relationships will be filled in after. He makes the point that either approach is good and that filling in the other set of relationships after works fine.
Another difference with PAMGO is that they they put in the top skeleton terms and will fill in the more detailed terms as annotators request them.
A URL for a cell cycle ontology has been circulated by someone. Chris is asked what he thinks of it.
He says it is not a new ontology and that for actual cell cycle terms it is identical to GO. It's just that they have added gene products and information from other sources. Its really an application ontology rather than a strict ontology.
He thinks there are no new relationship types but will take another look.
Status of OBO 1.2 release (relevant to content and software)
John is away but we are going to go ahead with the switch anyway, dependent on some testing.
We will test if obo2obo is working properly. We need to test if there will be a big diff. If all is well then Mike can set up a chron job and we can release.
After the release we will be editing the gene_ontology_edit.obo file and the gene_ontology.obo file will be automatically generated. It will be important that all editors should know this.
The users will not be affected.
action: Jane should mail mike and cc chris.
Cross product plugin in OBO-Edit now working. Cross products now live in sequence ontology.
We would like to do this in process ontology next. Will help with the long term names in cns terms.
We should discuss logistics.
We will do the internal cell type cross products first.
action: schedule chat. (who) Should include Karen Eilbeck.
There will be a new release soon.
AmiGO needs a big rework internally.
Chris asks if the working group will be happy with this. jane says that they understand that there are problems and that they will probably be happy with anything that seems likely to get things moving in the long run.
We need to find a way to get complaints from biologists so that we know what needs changed.
Jane is asked to find out what the biologists have been asking for by looking through the mailing list requests and other sources. Jane says she already knows all this.
Action item: Jane to get back to the management group with a list of the things that biologists request most that could be addressed with tools.
people want to do batch approaches and download in various formats. AmiGO is not like any individual mod. Seth said we could look at interface first as that can be kept same after rewrite. Chris likes Amelia's mock-up.
Suzi: In the reviews one of the comments was that GO software is too expensive to produce. Suzi is wondering about expanding the software group to include lots of other people round the consortium who write software and understand GO. Could include Amelia, Mary Dolan at MGI, Daniel Barrel at EBI, Sohel Merchant at DictyBase.
This is agreed.
The annotation outreach group have heard back from the PIs with a response to the priority list. Their main priority is to write SOPs for annotation submission by new groups. They are going to get together to make a plan after the Seattle meeting. Also Jen has got the database running and an ontology to hold information about collaborations. She is transferring the information from the text file to the database and ontology just now.
This group has made a lot of progress.
They have lists of genes through to October now. Rex is getting the list ready a month or two ahead so people have flexibility.
He is nervous about being dependent on Google spreadsheets but is going to set up a chron job to back up every few hours in case of accidents.
Suzi would like each of the reference genomes to provide three FASTA files to go on the GO website with all of a certain type of information for their species. There is a long discussion of the feasibility of this and the content of the files.
The file contents are to be approximately this, but more discussion is needed:
1) Gene models. Showing start to end of each protein non-coding RNA and any annotation that is available.
2) Protein set. For each genome, here are all the proteins.
3) FASTA file of transcipts. To include non-coding RNAs too.
Noted that Carol and Diana and Ewan working to reconcile ENSEMBL gene models (with something that I didn't get).
We could get the reconciled files and FASTA for proteins of gene models.
Action: Judy will talk to Karen.
Where will we put the files? Perhaps on GO site, perhaps just pointers to mod sites. Talk to Mike about this.
Action: this discussion has been taken offline and Suzi will head it up and discuss with the other PIs.
Back to reference genome annotation:
Rex has list of genes but need metrics too.
Need to formalize the format of the header for the files that the reference genomes groups submit. action: Mike will send format template.
Rex: people ask 'what genes can they curate?' but there is concern that curating orthologues only is limiting. Also how to track publications. This is more or less resolved. The goal is to get a ratio to show progress towards complete depth of annotation. Use total publications? Or triaged publications?
Either way is fine but has complications for the second column. You can use total publications associated with a gene. If no publications associated with the gene then that gene is fully annotated even when no annotation is done. The ratio must be valid.
Rex is concerned that if we choose a big group of genes to let people have flexibility then the focus and common goal outcome of the strategy will be lost. He is looking at 22 genes per month just now.
Also Rex would like to find a way for sourceforge items regarding reference genome genes to to be obvious to the reference genome curators. he wants everyone to be able to be involved in the discussions of strategy.
Considered using the 'group' function in the tracker facility. This would be difficult because the script that Midori wrote to generate the daily curator request tracker e-mail only reads the summary page.
There were other ideas:
Announcement on reference genome and annotation list. New reference genome curator list? These ideas were scratched because Midori's mail already goes to all GO list.
Jane: Why don't we just get people to put 'RG' at the beginning of the title of relevant sourceforge items. Then that would show up in the daily mail and no work would be involved in implementation.
(This idea seemed good but was not discussed. Perhaps people did not hear?)
Action item (all): Consider best way to solve this problem.
There is an AmiGO poster for the users meeting. Discussion of what to present at external advisory board meeting. Look-and-feel is important. Chris won't talk about that so Jane can.
Action: Jane will talk to Seth and Chris in Cambridge.
External Advisory Board Meeting, Sept. 7
Overview of presentations
goals from this morning's mail.
We are to talk about goals and what were going to deliver. Don't worry about dates.
The external advisory board will look at the deliverables that the whole group gives and help us to prioritize which are most important. Also bear in mind that we are planning for the next year and that it is a 5 year grant that we have applied for so we are not expecting everything to happen in the next month.
We can also ask for their ideas on how we can achieve what we want to achieve.
Suzi: what is online teaching in advocacy. That could be a deliverable. What do we want to teach? We could use same technology for teaching in all areas.
As a strategy we could make a list of possible tutorials. Rank in order of importance.
jane action: make this list for advocacy group.
later we will decide teaching medium.
Jane notes that lots of the information that we want to teach is hard to find.
Someone else mentions: A lot of what biologists want to do is impossible just now. Also many people don't want tutorials. What they want is the functionality in good intuitive tools.
Judy mentions that the GO Term Finder approach is good and intuitive. If AmiGO had that functionality incorporated then tutorials would not be necessary.
We should ask advisors about this. We need a user needs assessment (previous action item for Jane).
In reviewers comments they said survey was statistically unsound. They said it was good to do it but should be better written next time. Action: survey design could be carried out by the user advocacy group next time.
back to tools:
Do we want Amigo to do all the things that biologists want, or could we have several separate tools?
All needs to be in one tool to make things easy for the user.
Perhaps focus on microarray community as they are our main users.
Our database is all set up so would be much better if they come to our tool rather than have to download files and tools.
About ext adv board meeting.
Some comments on Midori's slides: Add more on why each method for developing terms is important. Not much background needed. They know what we do by now.
Put goals in presentations.
Judy: Mike can you do usage slides in case we need them? Ben is doing that. We will have that as a handout.
The cd has gone to the advisors.
We could bring 2 slides per page double sided or anything that we think the advisors would like to have.
Plan for the day. We are starting early.
There will be transportation from silver cloud.
OBO Anatomy, Sept. 8-9
Short summary: It should go ahead, though there are lots of problems.
GO Users (MGED9), Sept. 10
Few abstracts. Few people. Registration still open. Sent docs to mged people. Put abstracts on web, on go site.
Judy working on rollover money. Can't say until they get formal notification.
Summary of action items from this meeting
1. (Remaining from July 19) Will we use RCA for InterPro2go, TMHMM, etc? need input from Michelle Gwinn
2. Jane should mail mike and cc chris about switch to OBO 1.2
Test if obo2obo is working properly.
Test if there will be a big diff.
If all is well then Mike can set up a chron job and we can release.
Ensure editors know to switch to editing the new edit file.
3. Schedule chat about cross-products.
Invite Karen Eilbeck.
4. Jane to get back to the management group with a list of the things that biologists request most that could be addressed with tools.
5. Judy to talk to Karen about getting the reconciled ENSEMBL gene model files.
6. Discussion about three FASTA files to go on website is to be headed up by Suzi offline. To include PIs and Chris.
7. Mike will formalize the format of the header for the files that the reference genome groups submit and will send out.
8. We should all consider ways to let all reference genome curators know about any sourceforge items that relate to the gene lists.
9. Jane will talk to Seth and Chris in Cambridge about the new AmiGO mock-ups.
10. Jane will make a list of tutorials that are needed for the advocacy group. She will put them in order of importance.
11. Survey design to be statistically sound and to be carried out by the advocacy group next time.
Next meeting date
September 13th, 8 AM PDT, 10 AM CDT, 11 AM EDT, 4 PM BST
Back to minutes list