Itag Irc meeting 27-11-07
(01:55:34 PM) ThomasSchiex [i=93636fc2@gateway/web/cgi-irc/ircatwork.com/x-64d0a0803e8520e3] entered the room.
(01:56:02 PM) ThomasSchiex left the room (quit: Client Quit).
(01:56:06 PM) ThomasSchiex [i=93636fc2@gateway/web/cgi-irc/ircatwork.com/x-777dba3e18efc7e3] entered the room.
(01:56:31 PM) ThomasSchiex: Hello everybody
(01:57:02 PM) JanVanHaarst: Mark just left, he'll be back.
(01:59:21 PM) ThomasSchiex: Ok
(02:06:57 PM) dbuchan [n=dbuchan@ld-dbuchan.bioinformatics.ic.ac.uk] entered the room.
(02:07:03 PM) dbuchan: bonjour
(02:07:09 PM) dbuchan is now known as DanielBuchan
(02:35:30 PM) MarkF: Good, it appears that people from Ghent are able to join the meeting.
(02:35:37 PM) ThomasSchiex left the room (quit: "CGI:IRC (Ping timeout)").
(02:43:30 PM) ThomasSchiex [i=93636fc2@gateway/web/cgi-irc/ircatwork.com/x-fbdc5bec8b0ef5db] entered the room.
(02:45:04 PM) ErwinDatema [n=ErwinDat@chat.chabotc.nl] entered the room.
(02:45:30 PM) ErwinDatema: hi folks
(02:46:45 PM) DanielBuchan: Hi Erwin
(02:46:53 PM) MarkF: Hi everybody
(02:47:45 PM) ErwinDatema: We start in 10 minutes, right?
(02:48:16 PM) DanielBuchan: I think so
(02:48:30 PM) ThomasSchiex left the room (quit: "CGI:IRC (Ping timeout)").
(02:52:31 PM) MarkF: \W
(02:56:26 PM) MarkF: \W
(02:57:22 PM) ErwinDatema: .
(02:57:28 PM) MarkF: .?
(02:58:00 PM) JeffreyFawcett [i=9dc1caab@gateway/web/cgi-irc/ircatwork.com/x-c087eeef6487e44c] entered the room.
(02:58:45 PM) lukasmueller [i=84ec9d72@gateway/web/cgi-irc/ircatwork.com/x-01de7ccab4de3244] entered the room.
(02:58:56 PM) Anika [n=Anika@chat.chabotc.nl] entered the room.
(02:59:03 PM) lukasmueller: hello
(02:59:13 PM) Anika: Hello everybody
(02:59:17 PM) ErwinDatema: Hi all
(02:59:35 PM) lukasmueller: Rob can't join us today unfortunately... he is traveling..
(02:59:44 PM) JeffreyFawcett: hello everyone
(03:00:00 PM) lukasmueller: wow this seems to work!
(03:00:18 PM) ErwinDatema: Ah that's unfortunate, I wanted to discuss some pipeline related things with him, but I will email him about it
(03:00:30 PM) lukasmueller: Maybe I can answer them ....maybe not
(03:00:57 PM) MarkF: Hello
(03:01:10 PM) MarkF: Who are we still waiting for? Will Stephane join?
(03:01:17 PM) ErwinDatema: Well we had some problems with getting files from the repository using SCP here, I already discussed it with him a bit in some emails, and I wondered what his final solution would be
(03:01:45 PM) lukasmueller: Oh yes, what is that problem? I didn't quite get it from the emails I saw.
(03:02:29 PM) DanielBuchan: We always use sftp rather than scp, which I think gets round this problem
(03:02:31 PM) ThomasSchiex [i=93636fc2@gateway/web/cgi-irc/ircatwork.com/x-da789776b614a03c] entered the room.
(03:02:58 PM) lukasmueller: yes I think Rob also favoured sftp as a solution
(03:02:59 PM) MarkF: Or using a File of Filenames
(03:03:06 PM) ErwinDatema: Well basically, if you try to do a scp with wildcard in in (e.g., 'scp <sgn-server>:*fasta <our-own-server>'it will only retrieve 97 files at most
(03:03:26 PM) lukasmueller: I didn't know that scp had such a severe limitation...
(03:03:27 PM) ErwinDatema: but indeed, Rob mentioned to use sftp, which isn't really a problem either
(03:03:43 PM) lukasmueller: the problem is on the server side?
(03:03:49 PM) ErwinDatema: Well, not all versions do, but apprarently the one that you got on your machine does
(03:03:51 PM) strom [i=9dc1cb34@gateway/web/cgi-irc/ircatwork.com/x-335b05aa53583521] entered the room.
(03:03:52 PM) ErwinDatema: Yes
(03:04:11 PM) strom: Hi there!!
(03:04:29 PM) lukasmueller: Hi strom? Not sure who you are though
(03:04:39 PM) ErwinDatema: I sent a link to Rob with some onfo on it, he looked into it but the general conclusion was that it would be too much trouble to fix it
(03:04:44 PM) strom: sorry It's Stephane
(03:04:54 PM) ErwinDatema: *info
(03:04:55 PM) lukasmueller: Oh, hi Stephane!
(03:05:20 PM) DanielBuchan: It would require rebuilding your scp which would be a pain
(03:05:34 PM) lukasmueller: No, we don't want to do that...
(03:05:37 PM) ErwinDatema: Yes, let's not do that, I can use sftp
(03:05:54 PM) ErwinDatema: It just took me some time to figure out what exactly was going wrong
(03:05:59 PM) lukasmueller: What is the status of the pipeline? Have we run Eugene on all the batches?
(03:06:33 PM) ErwinDatema: I don't think so
(03:06:46 PM) lukasmueller: Where are the bottlenecks?
(03:06:56 PM) strom: No not yet... there is actually a little compilation problem with the very last version of EuGene... but I'm working on that right now
(03:06:58 PM) ErwinDatema: GenemarkHmm had an error, according to the email I got this morning
(03:07:19 PM) lukasmueller: GeneMark is running with Ath version?
(03:07:23 PM) ErwinDatema: I think Remy is responsible for running that
(03:07:33 PM) ThomasSchiex left the room (quit: "CGI:IRC (Ping timeout)").
(03:07:36 PM) lukasmueller: Or is Genemark calibrated?
(03:08:09 PM) JeffreyFawcett: I've run blastx and tblastx etc so that I can run eugene once the training is done and the latest version is properly installed
(03:08:12 PM) ErwinDatema: The analysis is called 'genemark_ath' so I assume it's Arabidopsis trained
(03:08:15 PM) strom: are any of those programs actually trained on tomato?
(03:08:20 PM) ThomasSchiex [i=93636fc2@gateway/web/cgi-irc/ircatwork.com/x-9857cf8bfd4aaf94] entered the room.
(03:08:23 PM) ErwinDatema: No, none are
(03:08:28 PM) DanielBuchan: We've solved all of our problems with Infernal so we can now process these batches inside of 24hours.
(03:08:35 PM) DanielBuchan: Where is the training set available?
(03:08:44 PM) Anika: In the EUSOL meeting in Rom there are some people who said they have tomato CDNAs or EST data. Could you integrate this data too?
(03:09:31 PM) DanielBuchan: Isn't the microtom cDNA data already integrated into the italian EST alignments?
(03:09:40 PM) ErwinDatema: I think it is
(03:10:00 PM) DanielBuchan: So that data is already integrated
(03:10:15 PM) remy__ [n=remy@chat.chabotc.nl] entered the room.
(03:10:18 PM) lukasmueller: There are a couple of gene finders that people would like to train on tomato and I would like to give them the training set. It would be done strictly in collaboration with ITAG.
(03:10:33 PM) lukasmueller: The gene finders are augustus and exon hunter
(03:10:43 PM) strom: Yes Marilu did has all the EST cDNAs around
(03:11:00 PM) ErwinDatema: Oh, that's great news Lukas!
(03:11:17 PM) ErwinDatema: I think the set that Stephane used is availabe on the iTAG wiki?
(03:11:23 PM) DanielBuchan: Where?
(03:11:40 PM) strom: yes on the training pages ...
(03:11:41 PM) lukasmueller: Yes, I got that reference... I hope everyone is ok with me giving it to these collaborators?
(03:11:52 PM) ErwinDatema: http://www.ab.wur.nl/mf/tomato/tomato.trainingset.tar.gz
(03:11:58 PM) DanielBuchan: Thank Erwin
(03:12:01 PM) DanielBuchan: ^thanks
(03:12:09 PM) ErwinDatema: It is linked from http://www.ab.wur.nl/TomatoWiki/TrainingSet
(03:12:23 PM) lukasmueller: So when can we run Eugene?
(03:12:30 PM) MarkF: Before using the trainingset, make sure you rebuild it!
(03:12:43 PM) ErwinDatema: Rebuild?
(03:12:47 PM) lukasmueller: what does that mean
(03:13:00 PM) strom: Lukas, we have to make sure that everything stays in iTAG.... and as we multiply the ab inito gene-callers we will have at some point to evaluate results...
(03:13:11 PM) MarkF: That means that the embl files are extracted from the wiki pages... it is not done on the fly.
(03:13:22 PM) ThomasSchiex left the room (quit: "CGI:IRC (Ping timeout)").
(03:13:49 PM) remy__ left the room (quit: kubrick.freenode.net irc.freenode.net).
(03:13:49 PM) Anika left the room (quit: kubrick.freenode.net irc.freenode.net).
(03:13:49 PM) JanVanHaarst left the room (quit: kubrick.freenode.net irc.freenode.net).
(03:13:49 PM) lukasmueller: Stephane, currently no gene finder seems to be calibrated... there are some very motivated individuals who want to help us...
(03:14:09 PM) ErwinDatema: Ah, I see the link to regenerate the archive, I will click it now so that we have the latest version of the archive rgiht now
(03:14:32 PM) MarkF: it needs to be regenerated every time somebody makes a change to any of the trainingset wiki pages.
(03:14:46 PM) TS2 [i=93636fc2@gateway/web/cgi-irc/ircatwork.com/x-54a3f1e07fe2a38e] entered the room.
(03:14:47 PM) MarkF: But don't regenerate to often, it takes quite some time.
(03:14:57 PM) ErwinDatema: Once Augustus has been trained for tomato we could integrate it into our own pipeline, to feed eEuGene
(03:15:30 PM) MarkF: What does training by the people from Augustus & Exonhunter mean? The models will become public, I assume?
(03:15:30 PM) lukasmueller: Yes, that's the idea. We'll probably run these additional gene finders at SGN or someone else can run them, and then feed into Eugene.
(03:15:40 PM) JanVanHaarst [n=JanVanHa@kokosnoot.wur.nl] entered the room.
(03:16:02 PM) MarkF: I don't see a problem with that
(03:16:37 PM) lukasmueller: The trained exon hunter will be made available to SGN. Probably the same with Augustus, but the "negotiations" are not that advanced quite yet. But Mario wants to help.
(03:16:43 PM) Remy__ [n=Remy@chat.chabotc.nl] entered the room.
(03:17:02 PM) strom: I know, and if we want to have something clean, we have to do the things right... so please do train the respective software for tomato as much as you can....
(03:17:36 PM) ErwinDatema: Well as long as they make the parameter files available to us, there should be no problem in them getting the training set?
(03:17:37 PM) DanielBuchan: genemark and glimmer really ought to be trained for tomato asap
(03:18:02 PM) MarkF: Lukas, what else will the people from Exon hunter & augustus do with the models? Will they become public?
(03:18:04 PM) Remy__: Genemark is almost ready, I would need some more gene models
(03:18:26 PM) JeffreyFawcett: I guess we're all working on that now, no? -including ourselves with the eugene ab-initio
(03:18:39 PM) ErwinDatema: I will look into training GlimmerHmm, I did it before with the previous training set and it wasn't that hard to train it, however the quality was -at the time- quite dissappointing
(03:18:43 PM) lukasmueller: As far as I understand, we will get the calibrated programs and we will run them here. So the models will be fed into Eugene. They could also be made public like any other results...
(03:19:46 PM) TS2 left the room (quit: "CGI:IRC (Ping timeout)").
(03:19:53 PM) TS3 [i=93636fc2@gateway/web/cgi-irc/ircatwork.com/x-a2a390c49af79a9c] entered the room.
(03:20:00 PM) Remy__: You want to make the matrices of the programs public?
(03:20:02 PM) TS3 left the room (quit: Client Quit).
(03:20:14 PM) Anika [n=Anika@chat.chabotc.nl] entered the room.
(03:20:14 PM) lukasmueller: It would be great to run the trained GlimmerHmm and then run Eugene based on the predictions if they are any good at all.
(03:20:19 PM) DanielBuchan: How many more bacs/model from the training set need checking?
(03:20:31 PM) strom: I would as much as possible send 1 iTAG annotation out, and avoid the overproduction of all kinds of predictions ... this is in nobodies interest
(03:20:44 PM) Anika left the room (quit: Read error: 145 (Connection timed out)).
(03:21:13 PM) Remy__: I agree with Stephane, too many programs can be confusing
(03:21:37 PM) lukasmueller: Stephane, I think your worries are premature... nobody else is releasing any predictions, it's all in a collaborative spirit with ITAG.
(03:22:04 PM) strom: users will want 1 prediction and all users will talk about that prediction .... any other will create a big mess
(03:23:22 PM) lukasmueller: Eugene can be fed with multiple gene predictors, that's what I'm talking about... nobody else will release another prediction.
(03:23:40 PM) MarkF left the room (quit: "User abort with 5 Ctrl-C's").
(03:23:54 PM) lukasmueller: We do need a number of trained programs for comparison... A comparison of just 1 or 2 programs is not meaningful.
(03:24:01 PM) strom: In that case it's OK and we will have 1 prediction that will go public
(03:24:10 PM) lukasmueller: yes
(03:24:14 PM) MarkF [n=mf@109-195.surfsnel.dsl.internl.net] entered the room.
(03:24:41 PM) ErwinDatema: Also, some users would also like to have access to tools with which they can run their own predictions on sequences that are not annotated by ITAG, e.g. BAC sequences from other solanaceae species, it would be nice to have such a tool available to thepose general public and I think a simple gene finder such as Augustus would be great for that pur
(03:25:10 PM) lukasmueller: I'm in the process of putting together a publication about the tomato genome initiative, but without any annotation it will not be interesting.
(03:25:26 PM) strom: what we need is an evaluation set... which no program/training ever saw... and on that set we will have to run each program individually, then via eugene
(03:25:55 PM) lukasmueller: We should generate a first, unofficial version of the protein set so that we can make some statements about the gene space etc.
(03:26:02 PM) lukasmueller: for the publication
(03:26:04 PM) Remy__: Stephane, won't you set up a web portal for BAC annotation?
(03:26:07 PM) MarkF: Does that mean we get a comparison between eugene and the other gene predictors? Or just eugene?
(03:26:31 PM) Remy__: We have to evaluate ALL gene finders independently
(03:26:47 PM) MarkF: But I agree with stephane, we need some quality assurance
(03:26:48 PM) Anika [n=Anika@chat.chabotc.nl] entered the room.
(03:26:53 PM) strom: Yes. for EU-SOL we have to setup a portal that would enable anybody to run eugene on there BAC (for other solanaceae)
(03:27:30 PM) DanielBuchan: But we don't really have enough HTGS3 sequence to derive an evaluation set. Right now it's better to use the sequence/models we have to train the genen finders
(03:28:01 PM) lukasmueller: we will also show all the individual results of the predictions, alignments and the final Eugene annotation on SGN.
(03:28:08 PM) MarkF: Stephane, will that be a web page? or a web service?
(03:28:14 PM) strom: and if we could get other 'ab initio' on the fly via some web-service that would be great... but that will make extensive blast etc not possible
(03:28:53 PM) MarkF: We will release a GlimmerHMM webservice.
(03:29:00 PM) MarkF: I'm already working on it.
(03:29:27 PM) strom: a web page + webservice calling other webservice if available
(03:29:47 PM) MarkF: Other gene predictorss would not be a big deal to set up too. I'm already working on GlimmerHMM, also from EUSol.
(03:30:04 PM) lukasmueller: these will all be trained versions?
(03:30:06 PM) MarkF: Stephane, what do you meanwith a webservice calling other webservice if available?
(03:30:11 PM) strom: will be OK for most 'ab initio'.... and that means that the 'ab initio' has to be good!!!
(03:30:23 PM) MarkF: If possible, yes. We will need to agree on that.
(03:31:25 PM) strom: if one asks for a prediction... we could ask (via webservice) an 'ab initio' results for 1 sequence and integrate the results in eugene
(03:31:55 PM) MarkF: That means that Eugene itself becomes available as a webservice?
(03:32:16 PM) strom: yes, as soon as we have a useble version
(03:32:22 PM) MarkF: Nice ;)
(03:32:35 PM) SchiexThomas [i=93636fc2@gateway/web/cgi-irc/ircatwork.com/x-11cb71169b6ecf7a] entered the room.
(03:32:47 PM) strom: and that is to help the potato and pepper people
(03:33:17 PM) lukasmueller: Cool... Speaking of the ITAG pipeline: When will these first contig batches be run through Eugene...? we need a timeline...
(03:34:08 PM) strom: So for the moment I try to get eugene3.5 running (have a library version problem) this is the version that read/write GFF3
(03:34:16 PM) MarkF: When will all predictors be trained?
(03:34:24 PM) ErwinDatema: Well there is an error in batch006 for genemark, I think Remy will have to fix that before EuGene can be run?
(03:34:25 PM) strom: it's a GD library
(03:34:37 PM) DanielBuchan: GD != fun
(03:34:54 PM) MarkF: GD? as part of eugene? I've fougt with it for Gbrowse... it's evil
(03:35:10 PM) Remy__: Yes, there is a problem with one bac, I'll fix it after the chat.
(03:35:17 PM) MarkF: Erwin, will you have some time to train Glimmer?
(03:35:26 PM) strom: Eugene has a static graph output too
(03:35:31 PM) ErwinDatema: No, but I will do it regardless
(03:35:37 PM) MarkF: good :)
(03:35:38 PM) lukasmueller: Great, so genemark is essentially ready... Does that mean we can run Eugene by the end of the week?
(03:35:58 PM) DanielBuchan: That would be excellent (especially so for our annual meeting in December)
(03:36:09 PM) lukasmueller: with both genemark and glimmer???? :_)
(03:36:10 PM) ErwinDatema: However, when we want to implement different versions of prediction tools, we will have to issue a nnew pipeline version!
(03:36:16 PM) lukasmueller: Yes
(03:36:20 PM) strom: second problem (and thomas is looking at that) is to have a better integration of splice site
(03:36:38 PM) strom: the score up to now were too close for neg. and pos.
(03:36:59 PM) MarkF: I think new versions of the pipeline are only essential when we move from pipeline version 000 to 001. This is still development.
(03:37:13 PM) ErwinDatema: annual meeting in December?
(03:37:18 PM) MarkF: Once everything is set up, we will need to rerun the complete pipeline anyway.
(03:37:24 PM) strom: but for the spice site them selves i have good results ... just have to make eugene deal with them better
(03:37:35 PM) lukasmueller: There is a SOL UK meeting on Dec 14- would be great to be able to show some results.
(03:37:35 PM) DanielBuchan: ErwinDatema: We have an annual UK sol research meeting in December
(03:37:50 PM) ErwinDatema: oh, UK meeting
(03:38:03 PM) SchiexThomas left the room (quit: "CGI:IRC (Ping timeout)").
(03:38:26 PM) lukasmueller: So Stephane, is it possible to run eugene by the end of the week???
(03:38:33 PM) strom: the whole pipeline doesn't need to be rerun .... all the blast will still be good enough
(03:38:56 PM) MarkF: Isn't the first question when Genemark & Glimmer are trained with the itag trainingset?
(03:39:01 PM) strom: end of this week... difficult... rather next week hopefully
(03:39:18 PM) lukasmueller: what is the timeline for glimmer Erwin?
(03:39:20 PM) strom: for the moment results are to weak...
(03:39:33 PM) MarkF: And does FGenesh need to be retrained with the new set?
(03:39:40 PM) MarkF: And GeneID?
(03:39:43 PM) lukasmueller: Yes, definitely...
(03:39:59 PM) Remy__: At the moment, the arabidopsis version of genemark is still better than the tomato version.
(03:40:00 PM) lukasmueller: Haven't heard from Francesco about geneid... good point.
(03:40:02 PM) strom: Francisco is working on GeneID
(03:40:05 PM) Anika: Maybe we should also integrate some CDNAs from potato till there are more sequences from tomato
(03:40:24 PM) ErwinDatema: The training itself doesn't take a lot of time, the problem will be tweaking the parameters in order to get the 'best' results.
(03:40:51 PM) strom: there are still a number of cDNAs out there ...
(03:40:57 PM) MarkF: Good, when do we think the 4 predictors are ready?
(03:41:09 PM) strom: what exactly do you need to train each soft???
(03:41:17 PM) MarkF: so that we can feed their predictions into eugene?
(03:41:39 PM) strom: couldn't we put that on the wiki.... everybody put out what he/she needs
(03:41:53 PM) MarkF: Lukas, what about FGenesh?
(03:42:39 PM) strom: if you can have it trained for tomato...
(03:42:52 PM) lukasmueller: FGenesh is being done by a company nobody knows what they are doing. They wouldn't use the training set anyway.
(03:43:08 PM) strom: but didn't you said it was not performing that well?
(03:43:14 PM) MarkF: ok, so we'll just have to live with that.
(03:43:16 PM) ErwinDatema: I am only here until the end of the week, after that I have holidays until the end of the year, so I will do my best to create a reasonably trained tomato version of GlimmerHMM before the end of the week, if that's not possible then I will have to find someone to take over for me
(03:43:26 PM) ErwinDatema: someone here in Wageningen, I mean
(03:43:39 PM) lukasmueller: I think nobody expects pipeline version 0 to produce perfect annotations... but it would be great to see that it actually runs and have some very preliminary results ...
(03:43:52 PM) ErwinDatema: Well we have that now
(03:44:12 PM) lukasmueller: For the latest batches? Where?
(03:44:19 PM) MarkF: Erwin, Bas or me will take over.
(03:44:23 PM) MarkF: if necessary
(03:44:26 PM) strom: but we have to aim as high as we can also (within reasonable time)
(03:44:43 PM) MarkF: Lukas, when do you really need the results?
(03:45:56 PM) lukasmueller: It would be good to have them by the end of the month... The intention was to submit the manuscript before year end.
(03:46:13 PM) ErwinDatema: End of the month is end of this week
(03:46:51 PM) lukasmueller: yes... Again, we'll only make summary statements in the ms... nobody will go and check individual exon borders... rough gene numbers etc should be +- ok from the eugene output
(03:47:05 PM) DanielBuchan: And it would be good for our UK annual meeting to have some preliminary results by the 14th
(03:47:16 PM) lukasmueller: agreed!
(03:47:36 PM) strom: I'm doing all I can to reach that .... but I also don't want to send out 'crap' :-)
(03:47:55 PM) strom: Can we come to an agreement:
(03:48:10 PM) lukasmueller: We're not sending out anything to anyone... These results are for internal use only...
(03:49:19 PM) Anika: Should the first functional annotation also be included?
(03:49:36 PM) Anika: In the paper..
(03:49:36 PM) lukasmueller: Yes... it would be great to have the interpro annotations in particular
(03:50:27 PM) strom: 1/ that all 'ab initio' get trained as much as possible... and evaluated if possible
(03:50:33 PM) DanielBuchan: The problem with this is that but interpro will take a long time to calculate (probably longer than a week) so the sooner we have some gene models the better
(03:51:01 PM) Anika: I could run all proteins through my pipeline, but this will also take some time...
(03:51:06 PM) MarkF: I need to go. I will assist Erwin if necessary. I will put a copy of this discussion on the wiki somewhere this week.
(03:51:11 PM) MarkF: cheers everybody..
(03:51:17 PM) lukasmueller: see ya!
(03:51:20 PM) strom: 1/ or we leave the other ab initio for after .... that would mean less training to be done
(03:51:41 PM) JeffreyFawcett: Can we leave integrating the other 'ab initio' gene finders till later on? - when all the gene finders we want to use are trained properly on tomato and ready?
(03:52:19 PM) DanielBuchan: I'd just use what the gene finders that are there can do now. More genefinders and better trainging can wait until a later iteration of the pipeline
(03:52:23 PM) lukasmueller: I think for version 0 we can work with trained glimmer and genemark, and run through Eugene. Annotate with interpro and Anika's pipeline. Get some initial data to evaluate the whole thing
(03:52:23 PM) ErwinDatema: I don't think it is realistic to do make changes to the pipeline now, if we want to finish before the end of the week - why don't we just run the pipeline as it is now on the data that we have?
(03:52:28 PM) strom: the EST + BLASTX + TBLASTX should do well already
(03:52:54 PM) JeffreyFawcett: that part is already done
(03:52:58 PM) lukasmueller: YES, Erwin, I agree with you =-)
(03:53:33 PM) DamnedIRC [i=93636fc2@gateway/web/cgi-irc/ircatwork.com/x-6ec49264e793a2e5] entered the room.
(03:53:57 PM) Anika: I also have to go, we have a meeting now. See you..
(03:53:58 PM) strom: doesn't need any change in the pipeline itself .... just not include specific files for the run of eugene ....
(03:54:07 PM) lukasmueller: We need to present something at UK SOL otherwise people will loose faith in ITAG... It doesn't have to be perfect, but we need to show that we can produce something.
(03:54:11 PM) DamnedIRC left the room (quit: Client Quit).
(03:54:29 PM) Anika left the room (quit: "Remote client closed connection").
(03:54:34 PM) lukasmueller: Of course we state that it is very preliminary...
(03:54:55 PM) ErwinDatema: So that means that from here, we have to run the EuGene predictions and after that the functional annotations? The rest seems to be done for batches 006 and 007
(03:55:23 PM) lukasmueller: Yes --- till the end of the week?
(03:55:51 PM) DanielBuchan: Importantly we have to show that there is some success. For the last 2 years for the annotation announcemetn at SOLUK we have said "there isn't enough data so can you please wait". We can't really say this again, not least as our funding body would not be impressed.
(03:56:02 PM) ErwinDatema: So that means Stephane has to get EuGene up and running and produce the integrated gene models, and from there they can be annotated
(03:56:15 PM) lukasmueller: Sorry I feel I'm a bit pushy... but I feel the same way as Daniel
(03:56:32 PM) strom: "crap in crap out"
(03:56:40 PM) DanielBuchan: sadly true
(03:56:41 PM) ErwinDatema: Daniel, if all else fails, you can always use the annotations from batches 1,2,3, which are ready
(03:56:49 PM) lukasmueller: better than nothing in nothing out...
(03:56:56 PM) DanielBuchan: ErwinDatema: Yes I would do that
(03:57:42 PM) Remy__: we really need to have the eugene output for all BACs/batches for the UK-Sol meeting!
(03:58:36 PM) lukasmueller: Remember that we need to resubmit a grant to NSF in January. I guess they would be more impressed with a rough approximation of a gene set than with nothing at all.
(03:59:54 PM) JeffreyFawcett: next week sounds a bit more realistic for us... ;-)
(04:00:24 PM) lukasmueller: Next week would still be great and there may be time to run some of the functional pipelines.
(04:00:27 PM) Remy__: beginning or end of the week?
(04:00:31 PM) JeffreyFawcett: when is the UK-Sol meeting daniel?
(04:00:40 PM) DanielBuchan: 14th of December
(04:00:40 PM) Remy__: 14th December
(04:00:56 PM) DanielBuchan: Next week would still be great
(04:01:22 PM) lukasmueller: should we have another meeting in a week with the idea that by then it is almost there?
(04:01:23 PM) strom: We'll to whatever we can to get a reasonable result that you can show ASAP
(04:01:33 PM) lukasmueller: Great!
(04:01:35 PM) DanielBuchan: That would be excellent.
(04:01:38 PM) DanielBuchan: thanks
(04:02:02 PM) strom: I was thinking of sending every day a little mail around so people would know
(04:02:26 PM) strom: and plan there part in accordance
(04:02:32 PM) lukasmueller: yes, that would be appreciated....
(04:02:47 PM) strom: so you'll see what is happening
(04:02:50 PM) lukasmueller: tentatively, we'll meet sometimes next week.
(04:02:55 PM) lukasmueller: right?
(04:03:11 PM) JeffreyFawcett: ok for me
(04:03:20 PM) strom: yes, but not next tuesday... won't be there
(04:03:27 PM) lukasmueller: wednesday
(04:03:28 PM) DanielBuchan: Well as this is an IRC room it could remain open all the time and we could arrange ad hoc discussions as an when necessary
(04:03:28 PM) lukasmueller: ?
(04:03:39 PM) lukasmueller: Yeah, great idea!
(04:03:39 PM) ErwinDatema: I won't be there myself but I'll tell Mark to be there
(04:03:49 PM) strom: OK
(04:03:59 PM) lukasmueller: Let's leave the chat open, but let's still schedual a meeting where everybody can participate
(04:04:09 PM) strom: YES
(04:04:10 PM) DanielBuchan: Wednesday next week is good for us
(04:04:25 PM) JeffreyFawcett: wednesday next week, same time?
(04:04:30 PM) Remy__: I think we should have a phone conference as the IRC takes too much time
(04:04:37 PM) lukasmueller: Let's say Wednesday then, yes, same time
(04:04:55 PM) strom: I agree .... also less discussions crossing
(04:05:09 PM) lukasmueller: Daniel, Sanger organizes a phone conference that works quite well, how about using that same system
(04:05:41 PM) lukasmueller: could you set that up, Daniel
(04:06:15 PM) DanielBuchan: it's the BT Pow Wow Now system, I think for most european countries you get charged local call rates. I'll look into it and email everyone
(04:06:46 PM) lukasmueller: I have a 1-800 number which is free I think... the conference organizer probably pays something...
(04:07:03 PM) DanielBuchan: Yeah we have a contract with BT at Imperial
(04:07:21 PM) lukasmueller: so that's perfect... right?
(04:07:50 PM) DanielBuchan: Yeah, I'll have to find out where they've hidden the details on the college intranet but as I'm aware it shouldn't be a problem at all.
(04:08:11 PM) strom: OK, keep us posted
(04:08:31 PM) lukasmueller: shall we call it a meeting... I'll leave the window open though
(04:09:07 PM) DanielBuchan: I'll book it for the same time next wednesday?
(04:09:08 PM) strom: could those that have an usable ab initio also keep everyone posted
(04:09:22 PM) lukasmueller: same time wednesday is ok
(04:09:27 PM) strom: yes
(04:09:28 PM) strom: ok
(04:09:45 PM) lukasmueller: bye!
(04:09:52 PM) DanielBuchan: bye lukas
(04:09:53 PM) Remy__: Let us have a meeting next Wednesday at 9am EST, either IRC or a real phone conference. Daniel, can you send out the details?
(04:09:54 PM) ErwinDatema: Daniel, could you please send around an email with the exact day and time on the mailing list?
(04:10:22 PM) DanielBuchan: Yes, once I've found out the details and booked it I'll email the itag list
(04:10:30 PM) strom: great
(04:10:32 PM) Remy__: thanks
(04:10:34 PM) Remy__: bye
(04:10:40 PM) JeffreyFawcett: see you guys!
(04:10:46 PM) strom: bye bye
(04:10:49 PM) ErwinDatema: ok, bye!
(04:11:11 PM) DanielBuchan left the room (quit: "using sirc version 2.211+KSIRC/1.3.11").
(04:11:35 PM) Remy__ left the room ("Leaving").
(04:18:14 PM) JeffreyFawcett left the room.
(04:41:02 PM) lukasmueller left the room (quit: "CGI:IRC (Ping timeout)").
(04:42:35 PM) strom left the room (quit: "CGI:IRC (Ping timeout)"). 