OE IRC 15June06
jrichter: Did everybody get a chance to look at the new technology demo?
J-Lo: i couldn't install GraphViz I'm afraid, so no
Harold: I've played with it
jrichter: Would the shifting around of the graph make it impossible to use?
tberardi: Just saw your email about 5 mins ago so, no.
Harold: I think it would be nice to have the "zoom" button that my graphviz has would be useful (I"m using the version that has been adapted to mac
jrichter: I haven't looked at that. It would be easy enough to add, but I couldn't figure out a way to zoom out and keep the node text readable.
jrichter: Is Midori planning to join?
MelissaH: Hi everyone.
jrichter: Hi, Melissa.
J-Lo: No, she's at the neuro meeting that Harold's not invited to
Harold: no, I don't think so; we had several emails and I wasn't even sure we would have anyone today.
jrichter: Without Midori, I'm guessing we don't have an official agenda.
tberardi: I thought Jane was going to be in charge.
Harold: John, My GraphViz settings for view things like interaction networks is set to "20" for the font size; that might be a way to start
J-Lo: I was supposed to chair - sorry, I thought the meeting was tomorrow - I have no agenda!
jrichter: Well, we can make this short and sweet.
Harold: david just ran by and said the neuroonto meeting agenda is running waybehind
MelissaH: I might need a little help getting your demo to load, perhaps we could all take a look at it?
tberardi: No joy with getting the Graphviz thing going on my PC. 8-(
jrichter: Let's see if there are other issues first, and then I can do a demo walkthrough with people who are interested.
kchris: I don't even have it downloaded yet, just got here...
jrichter: Did anyone have anything to say about new features, the annotation tool, etc?
tberardi: I was wondering why we hadn't heard from Suzi or Chris about the PATO annotation tool that's apparently being developed at Berkeley.
MelissaH: I haven't responded to the annotation tool yet, but
MelissaH: I was just going to mention that we'd been working with the berkeley group on this
kchris: Yea, I'd really like to know how any GO tool dev fits into the NCBO tool dev. It would really be silly to duplicate effort
tberardi: From talking to Dan Rubin, it sounded like this tool (Pheno?) could be used for GO annotation as well.
MelissaH: and that it does seem like any kind of annotation tool, whether it be for creating phenotype records, GO records, or expression records,
tberardi: Which, on the surface at least, sounds like what we're after.
MelissaH: is a very individualized thing. However, we would want to re-use parts of obo-edit for browsing and choosing terms.
kchris: It seems to me that each might different type of annotation might need it's own bit for what fields are needed, but that they could all work pretty much the same way
jrichter: Let me respond to some of what Melissa and Karen just said.
jrichter: 1) The Berkeley annotation tool may someday be usable for GO annotation, but my understanding is that, for now, it's really specific to PATO.
jrichter: 2) Any GOOD annotation tool is going to be highly specific to a particular group's needs. But at its heart, annotation is just associating some database record with some ontology term. I think it would be useful to provide a tool to do that very basic job and create an annotation file.
jrichter: This would mostly be of use to new or very small projects.
jrichter: (or research projects)
MelissaH: The PATO tool, or PHENOTE, allows user to choose genotype, or any other handle, and then terms from anatomy, GO, and PATO. It does seem like parts could be re-used for a more generic tool.
tberardi: Could the 'someday' in (1) be close to the same day that we develop a separate GO annotation tool?
jrichter: I don't know. I feel like 'someday' is at least 6 months away. Does that seem right to you, Melissa?
MelissaH: At ZFIN, we plan to integrate Phenote into our online curation tools. Flybase does it offline, so the tool was developed to accommidate both.
MelissaH: I'd say 6 months is about right.
MelissaH: thats when we will have phenotype support online.
jrichter: My main concern is that we don't duplicate anyone else's efforts.
MelissaH: It seems like the plugins to AmiGO or OBO-edit for any tool could be re-used.
jrichter: That doesn't mean that we shouldn't create a tool just because other tools exist; we just don't want to develop a tool that does EXACTLY what another tool does.
Harold: I kind of feel that the tool really has to be simple. It has to be easily installable with no IT support or even unix-y knowledge. It has to be for people that have little to no database experience. Something really along the lines of a 14 column spread sheet that can easily output a tab deleimted associatoon file.
jrichter: That's what I'm talking about.
Harold: or whatever the number of columns is
MelissaH: Its really the ability to connect to OBO-edit and then choose a term which populates the tool.
MelissaH: We curators here have no unix-y knowledge...
jrichter: Besides the OE working group, who else may be working on an annotation tool?
MelissaH: Its all happening behind the scenes with Java.
tberardi: Is Phenote that simple?
MelissaH: There is another evolution group that has made one similar to phenote for evolutionary phenotypes.
Harold: But you have a db; most users won't (to Mellisa). They at best will have a Genbank ID or UniProt id as the id to annotate to.
J-Lo: Lots of microarray tools have the ability to create an annotation file, but not GO format
MelissaH: I think thats ok, it would just spit it out as a gene association file rather than going into the database, right?
kchris: To be stand alone, it seems you'd need to be able to load in a lists of genes with IDs, as well as load in ontologies
MelissaH: But that doesn't mean it couldn't use a database for terms and IDs.
jrichter: Karen - We can use OBO instances for that.
kchris: John, what exactly does OBO instances mean?
jrichter: OBO format can do more than store ontology terms.
jrichter: It can also store data objects called instances.
jrichter: Technically, instances are the instantiated versions of ontology terms, but you can use instances for data storage if you build a special data storage ontology.
kchris: Would it possibly be simpler for users if the annotation tool (AT) can accept a tab delimited file for genes and IDs, i.e. something that could be kept in Excel?
jrichter: Sure. We could have it read in anything we want. The point is that since OBO-Edit supports everything in OBO format, and instances are in OBO format, we can represent our annotations using instances. So the infrastructure is already there.
MelissaH: Seems like this is something that the user would want to configure.
jrichter: Anyway, here's the point of my previous question - How can we find out that we aren't duplicating effort if we develop an annotation tool?
MelissaH: Send Mark to whomever we put in charge.
jrichter: Mark's the Phenote developer.
MelissaH: Phenote creator Mark Gibson.
MelissaH: We insisted he go to one of the evolution groups to make sure they were on the same page.
MelissaH: We could do the same for the GO-tool.
MelissaH: Then whomever the developer is could re-use parts of Phenote or at least coordinate efforts.
jrichter: I'm beginning to wonder if we need to discuss this at a consortium meeting. It seems that people have strong feelings about whether this should or should not be developed, and everyone is going to want a say.
MelissaH: I agree!
J-Lo: Don't we have an annotation tool WG now?
MelissaH: Identify user needs first.
jrichter: Maybe we should table this entire possibility until the next meeting (and maybe create a demo [or demos] to show at the meeting)
kchris: I think Alex's survey to go-friends is a really good idea to get a feel for what sort of tool might be useful, but beyond that, I do think that discussion at a GO meeting would be approrpriate
tberardi: Talk to the Phenote developer?
Harold: But that meeting is not until January now. I'm sure we can come up with something that more people could use immediately.
tberardi: (Whoops, lost track while Eva walked in and gave us a meeting update.)
jrichter: Harold - I agree, but there's lots of other things we could work on. Why spend our efforts on a tool that people could get really irritated about?
kchris: John, tabling for now, but Alex and the AT wg going ahead with the survey seems like a good start. We really shouldn't develop until we know what is wanted/needed
jrichter: Karen & Melissa - I agree that it's dumb to develop a tool that no one wants, but...
tberardi: Catching up, what meeting is not until Jan now?
jrichter: Sometimes I think it's better to offer a demo and then ask for responses than to ask users for their opinions with nothing to look at. Sometimes people don't think something is useful until it's right in front of them.
tberardi: John, how much effort would that involve?
Harold: the next GO consortium
MelissaH: true, but do we have any idea of how many groups don't currently have the capacity to create gene association files from their database?
kchris: I didn't say no one wants it. I just think that we should find out WHAT is wanted, and also get more info from the GOC, especially the PIs, about how this fits in with overall GOC goals.
Harold: But, Im sure ideas will come up in annotation camp?
jrichter: Tanya - depends on how functional it is. If it's just a mock-up of the interface, I could do it in 3 days. If it actually works, it could take 3 weeks.
jrichter: Karen - I agree.
jrichter: Okay, here's my motion -
MelissaH: Maybe you could help Mark integrate OBO-edit into Phenote as an example for everyone, many people need this already.
MelissaH: oops, sorry.
jrichter: I'd like to table this discussion until the next consortium meeting. But, I'd like to bring a working demo to that meeting to show the kind of thing this tool could do.
jrichter: However - If Alex's survey comes back with people screaming for a tool, we should revisit the discussion earlier.
tberardi: Would annotation camp (the first two days) be a good first opportunity to screen the demo, should it be developed? Or is that too soon? Are you coming to it, John?
jrichter: When is the annotation camp again?
tberardi: July 10-14
jrichter: There is a 75% chance I'll be there.
jrichter: But that's too soon for the annotation tool, I think.
kchris: You'd better let me know if you're coming, I need counts soon.
jrichter: Okay, Karen. I'll talk to Suzi tomorrow and see if I can get approval to go.
J-Lo: Okay - so the action item is
J-Lo: John makes a demo of OE annotation tool for jan consort meeting
jrichter: That's what I'd like to do, unless there are objections.
J-Lo: none here
kchris: Maybe I should make an agenda item for the Annotation Camp Part 1 to
kchris: talk a bit about this?
MelissaH: My only concern is that you coordinate with Mark so that the two tools can be designed similarly or can be integrated at some point.
jrichter: Melissa - I'd like to see what Mark's final design for Phenote is. Phenote is largely a user interface (as far as I know so far), so it may be that very little of Phenote will be usable in a Java tool.
MelissaH: One could imagine a whole tool package that allowed users to make many types of annotation.
jrichter: (Since interfaces always have to be redesigned when switching from a servlet architecture to Java)
MelissaH: its in java
jrichter: Sorry, I should have said "Swing" instead of Java.
jrichter: Anyway, I'll keep in touch with Mark.
MelissaH: nevertheless, I do think that if we are going to provide tools to new users, that they may want a variety of tools together.
jrichter: Karen - do you think an agenda item like that will be a good use of time?
MelissaH: just forward thinking.
tberardi: I also think that if the annotation tools are similar/the same, the learning curve will be that much smaller.
kchris: You know, what I really want regarding the potential GO annotation tool is some input from the GO PIs regarding where this fits into the grand plan, and we've had no joy from Suzi yet...
kchris: Maybe it would be better to jsut send an email to go-top to get clarification.
jrichter: Let's make that an action item. Any takers?
kchris: Maybe Alex would like to, as he's the leader of the AT WG, but I don't mind doing it.
kchris: I guess the action item could be I'll email Alex and we'll go from there.
jrichter: Alex's isn't actually here, he's just recording.
jrichter: Karen - I like that.
J-Lo: Okay - shall we move on?
kchris: OK, that's the action item. Let's move on.
J-Lo: We were going to discuss new features
J-Lo: but I think only Midori sent in her list of preferences so far
MelissaH: I love the select local button on the DAGviewer and would like to do the same on the GraphViewer.
jrichter: It seems that we've totally fallen off our deadlines for the feature preferences emails because of a flurry of travel.
kchris: Sorry, I really haven't had time to look through them yet, or add in an idea that I keep having when I'm away from the computer. Too much Annotation Camp stuff
jrichter: Can we push the deadline back to July 1st?
tberardi: I second.
MelissaH: yes please
J-Lo: lets do that.
tberardi: (Whoops, again. Just realized I'll be on vacation then. I'll email my prefs before I go.)
jrichter: Okay. I'll send out an email to the group saying that we're pushing back the deadline.
kchris: I'll be travelling from June 23 - July 3, but I'll try to get my prefs in before I go too.
jrichter: In the meantime, I'll work on small feature requests that I know we'll add and the graph demo.
J-Lo: can we add the relation filter thing that was demo-ed in the late OE betas?
J-Lo: it's really useful and already works...
jrichter: Yeah. If that's not on the feature request list, can you add it?
jrichter: I think we've got new features resolved then.
jrichter: I had a quick policy issue about betas I wanted to run by the working group...
jrichter: Before a new major release, obviously I'll release beta versions so we can work out the kinks in the new features. But when there's a bug report on an existing release, I've just been issuing a bug fix release without creating a beta for the bug fix. Is this acceptable?
tberardi: How do we know that the bug fix hasn't had a ripple effect without wg testing?
MelissaH: seems like a bug fix could create new bugs, right? then we wouldn't know why something stopped working.
kchris: Hmmm, maybe I'm just paranoid, but it seems like we should really test any version before it becomes the official version.
MelissaH: is there a way to put notes on a release, so we know it was changed?
tberardi: Unfortunately, this has happened all to often in my experience with bug fixes in PubSearch.
tberardi: 'too' not 'to'
jrichter: I see your point. But sometimes a bugfix is really tiny.
jrichter: (Like the last one)
MelissaH: I think its probably a judgement call on your part.
MelissaH: But also know that once we download a version, we don't go back and download it again with the fix unless we are told to.
kchris: It still seems like bad practice, it opens up the potential to designate an official version with some problem.
kchris: that we didn't catch
jrichter: Okay, I won't do that any more. All bug fixes require a beta release.
jrichter: How do we guarantee that these betas will get promoted to real releases quickly?
MelissaH: you prod us heavily.
tberardi: Set a designed testing time period and have designated people who will test out the basic features. (Responsibilities could rotate.)
kchris: Presumably these sorts of bug fixes are often things wanted by the Editorial Office, so I would expect that they will be very interested in doing the testing
kchris: and in poking other people to test. I also like Tanya's suggestion.
jrichter: Could we say that bug fix betas will be promoted to official releases in one week unless there is an objection from the WG?
MelissaH: The time might be different for different betas, though. some are big changes and some are small.
tberardi: I think there should be active confirmation from people who have tested and found that all is well.
kchris: I agree wtih Tanya
tberardi: Otherwise we run into the danger of nobody testing and promoting a version by default.
jrichter: That's exactly why I want a default promotion system. That danger will guarantee that betas get tested.
jrichter: (in a timely way).
kchris: Unfortunately, I don't think a default promotion system will guarantee that betas get tested.
MelissaH: I like Tanya's suggestion, you will need to designate testers.
jrichter: Okay - then we need to develop a process to approve betas.
tberardi: And they don't have to be the same ones all the time so that no one person is overloaded with testing duties.
kchris: Requiring a signoff will require that someone needs to do some poking to make sure that people do the testing, but maybe this responsibility can rotate
jrichter: How 'bout this:
jrichter: I'll do bug fix releases on Wednesdays. At the Thursday IRC meeting, we assign 3 (or more) people as testers. When those folks have signed off, we've got a release.
jrichter: The testers will be volunteers.
MelissaH: I like it.
jrichter: Sadly, this constitutes yet another incentive to skip these meetings.
J-Lo: we don't have these meetings very week though
J-Lo: volunteers do two releases?
MelissaH: testers could have a two week responsibility.
jrichter: That sounds good.
jrichter: And there won't be a beta release every week.
tberardi: Well, we could go for volunteers first and then, having exhausted all those who are at the meeting, designate absentees to do the work.
tberardi: And prod them heavily.
jrichter: Does that phrase make anyone else uncomfortable?
MelissaH: I live in farm country.
jrichter: We have 5 minutes. This really isn't enough time to walk through the demo.
Harold: We could run through it
MelissaH: how about at least telling us CS illiterates how to open it.
J-Lo: why will graphviz never work for me?
tberardi: I have another meeting but will stay for the next 5 mins.
jrichter: I don't know, J-Lo? Are you downloading the special mac version?
MelissaH: you probably don't have the path set right.
MelissaH: that seems to always be the problem here.
/Users/Applications/Graphviz 1.12 (v12)/Graphviz.app/Contents/MacOS/dot
J-Lo: "/Users/Applications/Graphviz 1.12 (v12)/Graphviz.app/Contents/MacOS/dot"
jrichter: Try renaming the folders so there are no spaces in the file names. That can sometimes cause a problem.
Harold: Jane, mine says "/Applications/Graphviz.app/Contents/MacOS/dot
jrichter: Would it be okay if we did another IRC tomorrow morning for people who need help walking through the demo?
MelissaH: Harold you are cut off on my screen.
J-Lo: I'll try that John
MelissaH: I can come.
Harold: Melissa: what do you mean?
kchris: ... and if you don't have Graphviz.app under Applications, where would you get it from?
MelissaH: I don't see the end of the path, only up to MacOS.
MelissaH: doesn't it end in dot?
jrichter: (for MacOS)
Harold: after MacOS/dot
Harold: yes, it ends "dot"
tberardi: Sorry, can't make that. But I'd be interested in reading the transcripts and following along at a later point.
jrichter: Anyway, I've got to go. I'll see interested parties tomorrow at 9:30 pacific time.
MelissaH: even if I make my screen bigger, long addresses get cut off.
MelissaH: bye bye.
kchris: I can't make tomorrow either, but following along from the transcript might help.
jrichter: Bye all. See you again soon.
MelissaH: Harold, can you email tanya the path?
MelissaH: I am on a PC.
MelissaH: ok, bye everyone.
J-Lo: I'm going to post the action items, even though everyone is off:
J-Lo: Add annotation tool to agenda for Jan consortium meeting - John to bring a working demo to that meeting to show the kind of things this tool could do
J-Lo: UNLESS Alex's survey comes back with people screaming for a tool, in which case we should revisit the discussion earlier.
J-Lo: Karen/Alex to find out from GO top where annotation tool fits in to grand plan
J-Lo: New policy: All bug fixes require a beta release.
J-Lo: Bug fix releases on Wednesdays. At the Thursday IRC meeting, we assign 3 (or more) people as testers. When those folks have signed off, we've got a release. Tester do two weeks, although there may not be a release every week
J-Lo: that's it
J-Lo: oops - I'd better save this transcript too
Harold: BTW, I might be joining the tool wg because it appears that I am outreach also