Camelia, the Perl 6 bug

IRC log for #bioperl, 2010-08-11

| Channels | #bioperl index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
01:48 CIA-95 bioperl-live: Jonathan "Duke" Leto topic/leto_interpro_bug * rdb77965 / (t/Ontology/Ontology.t t/data/interpro_sample.xml): Add test and test data for InterPro parsing bug - http://bit.ly/c2FpCK
08:33 dbolser deafferret: how hard would it be to implement gbrowse in Catalyst?
08:33 dbolser also, my fancy new reset button doesn't work
10:21 Yksi joined #bioperl
11:14 Yksi Is anyone here atm?
11:38 kai joined #bioperl
11:49 carandraug joined #bioperl
12:17 brandi joined #bioperl
12:17 brandi left #bioperl
12:21 dbolser no
12:21 dbolser ;-)
12:21 dbolser Yksi: irc is more like email than you think
12:23 Yksi dbolser:  It seems to vary greatly though. Some rooms, I've left my computer for about 10 minutes and I'm totally lost when i get back as conversation has moved on so far and i come back to 300 odd messages
12:28 dbolser right
12:28 dbolser try /names
12:28 dbolser and you can see howmany people are in the channel
12:28 dbolser this is roughly correlated to activity
12:28 dbolser channel activity
12:28 dbolser also, timezone of the /names will dictate busy periods
12:29 dbolser \names
12:29 dbolser ok, it is /names
12:33 Yksi The client that I'm using currently shows a list of names along the side. using the command /names does not give me any timezones sadly
12:35 dbolser no, unfortunately I think timezones are something you have to glean
12:35 Yksi mhmm
12:35 dbolser for example if you '/whois dbolser' you see the name of the box I'm on, which you can trace to the US, but I'm ssh'ed onto it from the UK
12:36 dbolser I'm trying to think of a humorous version of 'asl' that we can shout at people to mean something like, timezone, project, cs or bio, language of choice, msc / phd / postdoc / etc... ;-)
12:36 Yksi are you  from dundee
12:36 dbolser yes, how did you knwo :-)
12:37 Yksi simple google search of 'Dan Bolser - informatics'
12:37 Yksi leads me straight to your wiki page on bioinformatics.org
12:38 dbolser cool :-D
12:38 Yksi where did you do your masters?
12:38 Yksi (I'm doing *failing* mine at Exeter)
12:39 dbolser York
12:40 dbolser Don't worry, I didn't exactly pass either
12:40 dbolser what is the topic?
12:41 Yksi I've sailed through my exams and coursework and both the lit review and grant proposal.
12:41 Yksi my research project I just am failing totally at, and only have 3 weeks left
12:41 Yksi comparison of p. syringae strains b728a and ff5
12:42 dbolser sounds good
12:42 dbolser the project I mean
12:42 dbolser what is the grant proposal?
12:42 dbolser real or course work?
12:42 Yksi yes.. interesting but I just cannot do the bioperl
12:42 dbolser :-)
12:42 dbolser you need tools to get results that bp will help you handle
12:43 dbolser you have the genomes of both strains?
12:43 Yksi it was set as a coursework, but had to use the formal application methods
12:43 dbolser do you have to use bp?
12:43 dbolser cool
12:43 Yksi I've got the blast output… i just now have to write a bioperl script to parse the data
12:43 dbolser I wish I'd had that experience...
12:43 dbolser can you send me / post the exam questions?
12:43 Yksi for which?
12:43 dbolser for the grant proposal
12:43 Yksi Oh yeah
12:43 dbolser exam question / course material
12:44 dbolser :D
12:44 dbolser Yksi: for comparing similar strains, blast is not idea (IMHO)
12:44 dbolser have you tried mummer?
12:44 dbolser or a2a mapper # my current favourite
12:44 Yksi GAH !!! he told me to use blast to do this then bioperl to parse the data
12:44 * Yksi sighs
12:45 Yksi Yes i have a MUMmer output.. (i did that myself ages ago) then he said blast was necessary
12:45 dbolser well... he may be right... who is he?
12:45 dbolser oh
12:45 dbolser mummer should give you an 'overview' of the genome structure between the two domains
12:45 dbolser sorry, strains
12:45 Yksi well a 'Nucmer' output - table of the overal comparison of teh two strains
12:45 dbolser right
12:45 Yksi aligned seqs unaligned seqs stuff like that
12:45 Yksi *the not teh >.<
12:46 dbolser you should summarise the sequence diversity between the two strains by feeding your nucmer stuff into dnadiff
12:46 kblin Yksi: there's a blast parser in bioperl that's pretty decent
12:46 dbolser infact, dnadiff should give you 50 of your marks ;-)
12:46 Yksi How do i do dnadiff?
12:46 dbolser kblin: yeah, but summarising god knows how many hsps between two genomes into a project report... how would you do that?
12:47 dbolser Yksi: its part of the amos package that ships mummer
12:47 dbolser i think
12:47 kblin well, here's the part where bioinformatics gets tricky. you've got the raw data, now what? :)
12:48 dbolser right... which is why I'd go for nucmer or a2a mapper ... easier to interpret
12:48 dbolser or rather, easier to summarise ;-)
12:48 dbolser Yksi: are both genomes annotated?
12:48 dbolser cus you can prolly generate an annotation for both and do some comparative genomics
12:49 dbolser that would be at least 75% of the marks :-D
12:49 Yksi one is a fully published genome the other is not
12:49 Yksi i have the scaffolds and the contigs
12:49 dbolser also (tentatively) see #bioinformatics
12:49 dbolser it's a bit noisy in there, but you may get some help ;-)
12:49 kblin heh
12:50 kblin that's a way to put it
12:50 * dbolser is a part of the problem ;-)
12:51 Yksi I was also told to use geecee to generate histograms, though I am not really sure what purpose they serve
12:51 dbolser I don't know either
12:51 dbolser they are always there, but I don't know what they do...
12:51 * Yksi sighs
12:51 kblin Yksi: ok, I take p. is pseudomonas?
12:51 dbolser so Yksi, here is what you should do...
12:51 Yksi yes it is
12:52 dbolser 1) create a comparative genome assembly of strain x using genome of strain y (10%)
12:52 dbolser 2) create a comparative gene anotation of x using y and de-novo methods
12:52 dbolser 3) create a functional annotation of genes somehow (10%
12:52 dbolser (10% for 2 too)
12:52 dbolser compare genome structure
12:53 dbolser compare gene structure, including mutation rates and codon usage (30%)
12:53 Yksi dbolser:  sent you email with stuff
12:53 dbolser ty Yksi
12:53 dbolser get back to me tomorrow when you have done this much ;-)
12:53 Yksi O
12:53 Yksi :O
12:53 dbolser ;-)
12:53 Yksi i wont have done that by tomorrow :O
12:53 * Yksi dies
12:53 dbolser you can do it in a week I think...
12:54 dbolser the genome assembly will impress them
12:54 kblin Yksi: the gene annotation is doable
12:54 dbolser do that using minimus
12:54 dbolser do that using minimus2
12:54 dbolser again, part o famos
12:54 dbolser again, part of amos
12:54 Yksi I'm not sure what this amos is?!
12:55 dbolser https://sourceforge.net/apps/med​iawiki/amos/index.php?title=AMOS
12:55 kblin what's the gc content of pseudomonads?
12:55 dbolser https://sourceforge.net/apps/medi​awiki/amos/index.php?title=MUMmer
12:56 dbolser dnadiff is a part of mummer
12:57 Yksi dnadiff i think is what i've already run
12:57 Yksi you specify your reference sequence and your query sequence
12:57 Yksi and it outputs a table comparing the two?
12:57 nuba joined #bioperl
12:57 dbolser yes
12:57 Yksi I've done that (is included in the email)
12:58 dbolser oh, I thought you emailed me your grant proposal course work, not current course wokr
12:58 Yksi Ive emailed grant prop stuff too
12:58 dbolser Yksi: write that up ... write up the dna diff... but I'd do that after getting the genome assembled
12:58 dbolser ic ty
12:59 Yksi How am i going to generate the gene pools he wants me to?
12:59 dbolser you mean the list of genes in both genomes?
12:59 Yksi yes
13:00 Yksi and where they overlap
13:00 dbolser one should be downloadable
13:00 dbolser one should be generated by you using tool x
13:00 dbolser or method y
13:00 dbolser where method y includes a comparison
13:01 kblin you can probably use glimmer to predict ORFs
13:01 dbolser Yksi: take it one step at a time. you can do a lot in 3 weeks, but you need to 1) focus, 2) plan
13:01 dbolser or is it 1) plan and 2) focus... I always forget
13:01 dbolser but you get the point
13:01 dbolser ppl in here will help, but you shoudl stay logged in
13:03 dbolser wow... your masters looks awesome
13:03 Yksi I tend to stay logged in except for when i'm on the move or asleep
13:03 dbolser some dude gets you to write a grant for him and complains when you don't have the experience!
13:03 dbolser cool
13:03 Yksi (also having to send individual files because exeters mail system is wank)
13:04 dbolser cool
13:04 genehack dropbox++ # just sayin'
13:04 dbolser gmail++
13:04 genehack that too.
13:04 kblin tar?
13:05 dbolser boring!
13:05 Yksi i usually scp it to my web server and give urls to people.. but its a bit borked atm
13:05 kblin aw :(
13:05 Yksi [partner is in progress of switching servers around - thus im currently being hosted in two places which has somehow buggered up scp]
13:06 kblin he
13:06 kblin murphy's law says this stuff always happens at the most inopportune moment
13:06 kblin like the printer breaking on the night before you have to hand in your thesis
13:07 Yksi dbolser:  my masters is awesome! seriously the modules covered were really good! (taught masters) its just this research project.. I've actually gone from really wanting to work in the field to actually thinking 'oh crap have i got to go into more study to change field'
13:07 Yksi because i really jsut don't want to work with it anylonger :/
13:07 Yksi whcih is a shame after 4 years studying biology
13:07 Yksi and a nice hefty uni debt to go with it
13:08 kblin Yksi: but is it your field in general or just the research project?
13:08 kblin I mean, I really like microbiology
13:09 Yksi it's probably just my research project really :/
13:09 kblin but during my degree, I had to do stuff I couldn't care less about
13:09 Yksi my research project and phylogenetic trees …. if i have to do anymore of them i'll cry
13:09 kblin hehe
13:10 kblin for the latter there's a nice webserver in france somewhere
13:11 Yksi We were working on this work, and had been for about 4 weeks. and were using really old emacs to run this thing as it was the only supported machine…. It had been running for 6 days and was nearly complete… and some moron fused the rooms electricity supply… everything went off
13:11 Yksi i lost all that work
13:11 Yksi about 4 days before hand-in
13:12 Yksi of course not leaving me enough time to re-run it properly… I used a really rubbish web one that could run and bootstrap in about half an hour. Needless to say, it was my lowest marked piece of work
13:14 Yksi so am i totally going off-track still trying to write this bioperl script to parse the data?
13:15 kblin ok, what do you want to do with the data once it's parsed?
13:21 Yksi okay from last times meeting…. I've apparently got to parse the data, then create an array for the aligned genes and an array for unaligned geens of each pathavar. I've got to generate a gene pool . and use CG view to generate visible outputs
13:24 kblin look at Bio::SearchIO for parsing the blast results
13:25 Yksi kblin: I have been :p
13:36 dbolser I never touched another tree since my masters
13:36 Yksi dbolser: that would be a good thing :P
13:37 _ilbot2 joined #bioperl
13:37 Topic for #bioperlis now http://bioperl.org/wiki/IRC#Getting_help | http://www.bioperl.org/wiki/Using_Git | paste to http://pastebin.com/
14:17 deafferret ..o.O
14:17 kblin I've touched phylogenetic trees for colleagues
14:18 kblin using http://www.phylogeny.fr/
14:20 Yksi I've got one in my project, but its only a reference one and i dont actually have to do anything to it :P
14:25 Yksi gah I've ground to a halt today. brainn has given up on me
14:27 deafferret MOAR COFFEE
14:28 Yksi haha i wish :P i don't like coffee
14:28 deafferret THAT WASN'T A QUESTION, PRIVATE!! DRINK!
14:28 Yksi nuuuu
14:30 kblin you could start the genome assembly
14:30 kblin then your computer would be busy and you could take the rest of the day off
14:31 Yksi takes that long?
14:35 Yksi and i use AMOS to do it?
14:35 Yksi (its not something i've heard of prior to today)
14:42 deafferret start it on all my computers please
14:44 kblin Yksi: depends, how big is a pseudomonas genome? and how slow is your computer?
14:44 Yksi if i can get it instaled on my mac - pretty quick
14:44 Yksi own server - pretty slow
14:47 Yksi the ref sequence (B728a) has 6093698 bases (according to mummer) and the query ( ff5  - i.e one in contig/scaff format) is 5656753 basess
14:55 dbolser Yksi: how many pieces each?
14:56 Yksi err no idea :S
14:56 kblin deafferret: http://xkcd.com/303/ ;)
14:56 Yksi XD thats awesome
14:56 deafferret or "assembling", for bioinformatics  :)
14:57 deafferret takes WAY longer than compiling  :)
14:57 kblin same general idea
14:59 deafferret on 2nd thought, those assemblers are getting pretty damn fast...
15:00 deafferret "running MY assembler" should always buy a couple weeks
15:00 Yksi i hope mine wont take that long :S
15:08 kblin speaking of xkcd, http://xkcd.com/765/ I especially like the alt text
15:10 Yksi http://xkcd.com/705/ i like this one… its just what my partner is like
17:48 dukeleto anybody know what is up with this? http://gist.github.com/519384
17:48 dukeleto i made a topic branch for another interpro parsing bug that I found, topic/leto_interpro_bug
17:49 pyrimidine joined #bioperl
17:49 dukeleto pyrimidine: hey there!
17:49 pyrimidine o/
17:49 dukeleto pyrimidine: i wouldn't mind if you could check out http://gist.github.com/519384
17:50 dukeleto pyrimidine: there are more interpro parser bugs lurking
17:50 pyrimidine wouldn't be too surprised at that, unfortunately
17:51 pyrimidine is this on a branch?
17:52 pyrimidine dukeleto: ^^^
17:52 dukeleto pyrimidine: yes, , topic/leto_interpro_bug
17:52 pyrimidine ok
17:54 dukeleto pyrimidine: the issue is in Bio/OntologyIO/Handlers/InterProHandler.pm
17:55 dukeleto pyrimidine: the if( ! defined $term_temp ) adds a term and then marks it uninstantiated, but something is not right, because when the actual term gets loaded, it thinks it is already there
17:56 pyrimidine weird, the stack trace was cut off
17:56 dukeleto pyrimidine: the new test file has a term that is the parent of terms that come before it in the file
17:57 dukeleto pyrimidine: so a placeholder term is added for it, but then things blow up when the actual term is seen
17:57 pyrimidine yep
17:57 dnewkirk joined #bioperl
17:58 pyrimidine so, it seems there should be a check for the presence of that term
17:58 dukeleto pyrimidine: that is what I was thinking, but there seems to be something wrong with mark_uninstantiated
17:59 dukeleto pyrimidine: only a placeholder of the term is added, with just a name and identifier. all the data isn't there, because it hasn't been seen by the parser yet
17:59 dukeleto pyrimidine: so checking if it already exists wouldn't fix this, i don't think
17:59 * deafferret does his funny in - ter - PRO! voice
18:00 pyrimidine no, probably won't
18:04 pyrimidine does seem there is something wrong in mark_uninstantiated
18:05 pyrimidine or there isn't a check for that in add_term
18:08 pyrimidine looks like there isn't a check for presence of the term in the instantiated term store
18:10 dukeleto mark_uninstantiated probably needs some tests
18:13 pyrimidine yes
18:14 CIA-95 bioperl-live: Chris Fields topic/leto_interpro_bug * r1fa8390 / Bio/Ontology/SimpleOntologyEngine.pm : check for _instantiated-edness - http://bit.ly/9ltjQf
18:14 pyrimidine try that one
18:14 pyrimidine though it's getting some warnings
18:14 pyrimidine it's passing
18:15 pyrimidine dukeleto: ^^^
18:18 pyrimidine afk # coffee!!
18:19 CIA-95 bioperl-live: Chris Fields topic/leto_interpro_bug * raa9bde1 / Bio/Ontology/SimpleOntologyEngine.pm : ne, not != - http://bit.ly/a7MuZw
19:06 newtobioperl joined #bioperl
19:08 newtobioperl HI everyone. I have a question. Lets say I have:  chromosome number, start-end position of that chromosome (human). Is there any way to know if there is a coding region (exon, intron) of that sequence (between start-end)???? thank you
19:11 deafferret well, step 1 is to pick a datasource.
19:11 deafferret lately I've been doing raw EMBL files for human for that sort of thing
19:11 deafferret the Ensembl and UCSC browsers will show you dozens of tracks of information, and you can interact with their databases I think
19:35 newtobioperl thanks. What if i have about 90000 sequences like that. Can I use bioperl to get that information from NCBI or UCSC?
19:37 newtobioperl Basically, i just want to see if there is a coding region between those start-end position
19:37 newtobioperl and somehow print out a visualize report
19:44 rbuels deafferret: ^^
19:44 * rbuels would help, but knows pretty much nothing about human data sources
19:44 deafferret rbuels: ^^
19:44 CIA-95 bioperl-live: Scott Cain master * r7218728 / Bio/FeatureIO/gff.pm :
19:44 CIA-95 bioperl-live: tiny change to Bio::FeatureIO::gff to allow the gmod chado gff3 bulk loader
19:44 CIA-95 bioperl-live: to not choke when the gff file has ##sequence-region directives. The loader
19:44 CIA-95 bioperl-live: is documented not to support this, but now it will quitely ignore those
19:44 CIA-95 bioperl-live: directives. - http://bit.ly/cA31Gc
19:45 rbuels dukeleto: ^^
19:45 newtobioperl :)
19:45 rbuels "the loader is documented not to support this", lol
19:45 deafferret newtobioperl: pick a sequence. pick a datasource (UCSC?). figure out which of 100 tracks you care about via their web browser. then figure out how to dynamically find the same data
19:46 deafferret then get banned cause you're not supposed to hit them 90,000 time
19:46 deafferret s
19:46 deafferret but at least then you'll have a feel for what data exists and what data your customers care about
19:46 deafferret then tell them "everything" is too vague
19:47 deafferret and force them to make a few decisions to start your journey
19:47 rbuels biology is too vague.
19:47 newtobioperl lol
19:47 deafferret after a few years, become frustrated and jaded and wander off and start your own basket weaving company
19:48 newtobioperl I am thinking about using Gbrowse to print out the visualized results
19:48 deafferret :)
19:48 deafferret umm.... you need to figure out what data you care about before you start trying to build your own database and visualizers
19:49 newtobioperl i have been using UCSC human genome
19:49 deafferret k. so do you know how to get (via the browser) exactly the tracks you want?
19:49 newtobioperl and gbrowse for visualized reports., Now i want to go a little bit deeper
19:50 newtobioperl I mean GS mapper give me the start-end position + chromosome #
19:50 newtobioperl isnt that enough?
19:50 deafferret you're the judge of what's "enough" not me  :)
19:50 deafferret what is the "GS mapper" ?
19:51 newtobioperl 454 GS Mapper
19:51 newtobioperl its from Roche
19:51 deafferret i have zero experience with 454 software
19:53 newtobioperl well, basically their results are like blast
19:53 * rbuels wonders about software written by roaches
19:53 newtobioperl its very fast though
19:54 newtobioperl I use GS Mapper and run my 94k queries against UCSC human genome.
19:54 newtobioperl It takes about 15 mins
19:54 newtobioperl 94000 sequences*
19:55 deafferret what tool(s) do you use for the UCSC side of that?
19:55 deafferret URLs?
19:55 deafferret mirrors?
19:56 newtobioperl it comes in a package when you buy their computer
19:57 newtobioperl http://454.com/
19:57 deafferret 454 comes with a copy of UCSC stuff?
19:57 newtobioperl oh sorry
19:57 newtobioperl what do you mean UCSC side?
19:57 newtobioperl I use Gbrowse
19:57 newtobioperl for visualized results
19:57 newtobioperl GS Mapper --> parser --> GFF >>Gbrowse
19:58 deafferret tell me more about "and run my 94k queries against UCSC human genome"
19:58 newtobioperl k
19:58 deafferret are you just mapping your reads to that reference assembly?
19:58 newtobioperl in my lab
19:58 newtobioperl we are trying to analyze HIV data
19:59 newtobioperl we want to know the locations of HIV where they integrate themselves into human genome
19:59 newtobioperl we divide them into 2 cat: Activated and unactivated
19:59 deafferret are you just mapping your reads to that reference assembly?
19:59 newtobioperl after we trim off the HIV genes
20:00 newtobioperl we got 94k sequences for Activated
20:00 newtobioperl yes
20:00 newtobioperl then I mapped those 94k sequences to UCSC human genome
20:00 newtobioperl using GS Mapper
20:01 newtobioperl then I parse the results to Gbrowse
20:01 deafferret ok. so I'm now under the impression that you just have the assembly sequence from UCSC.   but if your next goal is to map all the proteins or other features, then you need to dive FAR deeper into the UCSC dataset
20:01 deafferret they have hundreds of tracks of information at UCSC
20:01 deafferret you're only using the sequence
20:01 newtobioperl yes
20:01 newtobioperl i downloaded 23 chromosomes from UCSC
20:02 deafferret k. so, go to their website and decide what tracks you care about...
20:02 deafferret in what format?
20:02 deafferret Primus++
20:04 newtobioperl i am trying to upload my own GFF to see what happens
20:05 newtobioperl what is primus++?
20:07 deafferret http://en.wikipedia.org/wiki/Primus_%28band%29   :)
20:09 newtobioperl ... haha
20:09 newtobioperl well
20:09 deafferret http://www.primusville.com/
20:09 deafferret i need me one of those   :)
20:10 newtobioperl hahah
20:10 newtobioperl well, I can't do this with 94k results.
20:11 newtobioperl is there another approach for this deafferret?
20:13 newtobioperl I just want to know if there is coding region(s) between those start-end position. Forget about visualize results for now
20:13 newtobioperl and how can I do it with about 100k sequences
20:14 deafferret i'm begging you to click around in the UCSC website and figure out which tracks you care about
20:14 deafferret there are many flavors of "coding region" as far as I can tell
20:17 deafferret http://genome.ucsc.edu/cgi-bin/hgTra​cks?hgTracksConfigPage=notSet?org=Hu​man&amp;db=hg19&amp;hgsid=167368978   "expand all"
20:17 deafferret "Genes and Gene Prediction Tracks"
20:18 deafferret as you can see, "coding region" means different things to different people
20:19 deafferret lately, we've blindly settled on building our own SQLite database from raw EMBL files + BioPerl
20:19 deafferret i pluck out the 5', 3' UTRs and regex subseq match them.... etc
20:20 deafferret (  ... has anyone else noticed that if a bioinformatics question is possible to answer that's only because the question is too vague? )
20:23 Yksi1 joined #bioperl
20:24 newtobioperl Quick question: What does the blue text mean on this page? http://genome.ucsc.edu/cgi-bin/hgc?g=ht​cCdnaAli&amp;i=U03241&amp;c=chr21&amp;l​=19995140&amp;r=21145303&amp;o=20289530​&amp;aliTable=all_mrna&amp;table=mrna
20:25 deafferret "Matching bases in cDNA and genomic sequences are colored blue and capitalized"
20:27 newtobioperl thank you
20:28 Yksi joined #bioperl
20:38 dukeleto pyrimidine: looks like you fixed that interpro bug, nice! Do we need more tests, or is it ready to merge?
20:39 pyrimidine dukeleto: It can be merged, I think your test covers it
20:39 pyrimidine passes all other tests as well
20:44 dukeleto pyrimidine: cool, will do
20:58 CIA-95 bioperl-live: Jonathan "Duke" Leto master * r30616be / (t/Ontology/Ontology.t t/data/interpro_sample.xml): Add test and test data for InterPro parsing bug - http://bit.ly/aQHVfF
20:58 CIA-95 bioperl-live: Chris Fields master * r27ca68b / Bio/Ontology/SimpleOntologyEngine.pm : check for _instantiated-edness - http://bit.ly/bNnIuc
20:58 CIA-95 bioperl-live: Chris Fields master * r5ae510c / Bio/Ontology/SimpleOntologyEngine.pm : ne, not != - http://bit.ly/bMyZMh
21:08 dukeleto pyrimidine: more bugs. now I get a "relationship already defined" error with the full InterPro XML
21:10 dukeleto pyrimidine: the exception dumps the contents of each variable and ends up being 20K lines of output. classy.
21:11 pyrimidine nice!
21:12 pyrimidine may have to file that one.
21:13 pyrimidine (or at least whittle down the data somewhat)
21:15 dukeleto pyrimidine: yeah, i will attempt to create a small test file. It really sucks whittling down a 76MB xml file, it is so slow to edit
21:15 dukeleto pyrimidine: is there some way to mark a relationship as uninstantiated, or does that not make sense?
21:16 pyrimidine not sure, unfortunately I didn't write that code
21:16 deafferret slow to edit in what?
21:16 * deafferret uses vi or XML::Twig for everything XML
21:17 pyrimidine The bioperl way of doing it: look for whatever XML parser is being used in other modules
21:17 pyrimidine then use something differernt
21:17 dukeleto deafferret: in vim. guess i should turn syntax highlighting off..
21:17 pyrimidine *different
21:26 CIA-95 bioperl-live: Jonathan "Duke" Leto topic/leto_interpro_relationship * r83f6ec5 / (t/Ontology/Ontology.t t/data/interpro_relationship.xml): Add a test and data for a bug in InterPro relationships - http://bit.ly/9zmOxw
21:27 dukeleto pyrimidine: there ya go, a nice 20k line failing test output :)
21:29 dukeleto pyrimidine: the important line of that output is: between IPR000086 and IPR000059 already defined as
21:30 dukeleto pyrimidine: it may have something to do with the interpro file having redundant relationships, such as saying "a is a parent of b" and then later on "b is a child of a"
21:30 dukeleto pyrimidine: not sure, tho
21:35 pyrimidine dukeleto: will look at it
21:36 genehack deafferret++ # just caught up on the impromptu bioinformatics practicum
21:36 deafferret heh
21:36 genehack you, sir, are a patient man.
21:37 deafferret oh? i read me as generally grumpy today
21:38 deafferret Newsflash: IE sucks.  http://www.ejeliot.com/blog/63   IE version specific CSS tags?!? are you f'n kidding me??
21:38 deafferret HOLY CRAP. I miss Monday where I spent all day fighting lpr queues
21:39 * deafferret reminisces
21:39 genehack OT but hiiiilarious: http://whatthefuckismyinfor​mationsecuritystrategy.com/
21:41 deafferret ooo... looks like the titles of most IS papers
21:41 deafferret that is handy!
21:42 pyrimidine dukeleto: my guess is you are correct (repeated relationship).  That one may be more difficult.  Personally, seems like a bit of a kludge to have one list of terms/relationships that is instantiated, others that aren't.
21:46 bag_ joined #bioperl
21:56 nuba joined #bioperl
23:48 Yksi joined #bioperl

| Channels | #bioperl index | Today | | Search | Google Search | Plain-Text | summary