Camelia, the Perl 6 bug

IRC log for #bioperl, 2009-06-03

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

All times shown according to UTC.

Time Nick Message
00:11 alper_ joined #bioperl
11:34 perl_splut joined #bioperl
12:06 kyanardag joined #bioperl
12:48 faceface what time is it in the US?
12:50 pyrimidine joined #bioperl
13:49 deafferret faceface: 0849 here
13:49 deafferret central US
13:49 balin joined #bioperl
13:50 balin_ joined #bioperl
14:05 faceface tyty
14:12 deafferret faceface: you're welcome. I'm happy to be much slower and less accurate than a google search  :)
15:28 perl_splut mainland US is GMT-5 to GMT-8
16:09 faceface tyty
16:09 faceface whats the deal about solexa changing the quality scale for there 'phrap like' scores?
16:09 faceface where is teh difinitive resource for this discussion?
16:10 faceface (seemed to come up about 5 times on the mailing list recently)
16:12 rbuels faceface: you're looking for a definitive resource?  in bioinformatics?
16:13 faceface rbuels: or a bunch of vague resources, and I'll write the definitive
16:13 * rbuels doesn't know anything about solexa qual anyway
16:22 perl_splut wonder if it would be worthwhile to create some modules to automate SpectrumMill
16:23 * deafferret would be happy to sell you whatever definitiveness you'd like to pay for   :)
16:32 pyrimidine joined #bioperl
16:34 faceface It also means that there are now three types of FASTQ files floating about; standard Sanger FASTQ with quality scores expressed as ASCII(Qphred+33), Solexa FASTQ with ASCII(Qsolexa+64) and Solexa FASTQ with ASCII(Qphred+64).
16:35 faceface http://seqanswers.com/forums/showthread.php?t=1110
16:36 faceface deafferret: I'll pay for 95% certainty
16:40 pyrimidine Wikipedia is sometimes your friend: http://en.wikipedia.org/wiki/FASTQ_format
16:40 pyrimidine (and yes it's a mess)
16:41 * rbuels slaps forehead
16:41 rbuels i knew about that, should have told faceface
16:41 pyrimidine Also: http://maq.sourceforge.net/fastq.shtml
16:41 * faceface slaps rbuels forhead
16:41 faceface ;-)
16:41 rbuels lol
16:41 faceface ty pyrimidine
16:42 * rbuels now has a big red mark on his forehead
16:43 * faceface gets out his universal color chart
16:43 rbuels i think it's about #ff2a2a
16:44 * pyrimidine thinking Heikki was supposed to add support to bp
16:44 pyrimidine that being for solexa FASTQ
16:45 rbuels pyrimidine: one of the tomato genome annotation pipelines has a bunch of bio::seqfeature::annotated-based stuff in it.  bsfa is deprecated or something now right?
16:45 pyrimidine if I have anything to do with it, yes
16:45 pyrimidine it doesn't do what's it's supposed to do very well
16:46 rbuels the better thing is typedseqfeaturei and ...something
16:46 rbuels right?
16:46 pyrimidine yes
16:46 deafferret laugh
16:46 * rbuels slaps deafferret's forehead
16:46 rbuels lol
16:46 deafferret woo woowoowoowoowoo!
16:47 deafferret nyuck nyuck
16:47 rbuels deafferret: where are you staying for yapc, you get a dorm room?
16:47 pyrimidine well, it shouldn't be hard to implement something that is typedseqfeaturei
16:47 rbuels (you knucklehead)
16:47 deafferret ya
16:47 rbuels pyrimidine: i implemented an fgenesh and genomethreader parser using bsfa
16:47 pyrimidine use base qw(Bio::Root::Root Bio::SeqFeature::TypedSeqFeatureI);
16:47 rbuels pyrimidine: and i wanna fix them
16:47 pyrimidine Here's the specific issue:
16:48 pyrimidine BSFA does everything on the fly and types everything
16:48 pyrimidine (literally, every bit of data)
16:48 pyrimidine into something along the lines of Bio::AnnotationI
16:48 pyrimidine It's a little heavy-handed
16:49 pyrimidine (and here's the best part)
16:49 philsf joined #bioperl
16:50 pyrimidine it downloads so.obo
16:50 pyrimidine (the entire thing)
16:50 pyrimidine to type your data.
16:50 pyrimidine This should be done factory-style
16:51 pyrimidine http://bugzilla.open-bio.org/show_bug.cgi?id=2513
16:51 pyrimidine for those who are interested
16:51 pyrimidine Here's the catch:
16:52 pyrimidine If you have an entire Ontology loaded (class-based attribute BTW), and you try to persist it via Storable
16:52 pyrimidine (as Bio::DB::SeqFeature::Store does)
16:53 pyrimidine then you end up storing a copy of the entire Ontology object along with each BSFA instance
16:53 rbuels yarp.
16:53 rbuels heh.
16:53 pyrimidine yarp is right
16:53 rbuels so what's this TypeFactory idea you have
16:54 pyrimidine Well, the factory would do the heavy lifting
16:54 pyrimidine and switch the objects passed through from an untyped SF to a typed SF
16:54 pyrimidine (possibly unflatten, do heavy typing, etc)
16:55 pyrimidine in this case, the Ontology would be in the factory, not the feature, so it's much lighter
16:55 pyrimidine (and we don't have to muck around with typing everything if we don't want to)
16:55 rbuels so the primary_tag() of each feature would be SO:0023353 or whatever?
16:55 rbuels (for a typed seqfeature?)
16:55 pyrimidine yes
16:56 pyrimidine in theory, this should be easy to set up
16:56 rbuels ok i can see that.  and a typed feature would have a method to get the ontology term obj for its primary_tag()
16:56 pyrimidine yes
16:56 rbuels via the .... typefactory
16:56 rbuels ?
16:56 pyrimidine (note that there is a Bio::Annotation::OntologyTerm)
16:56 pyrimidine yes
16:57 rbuels is b:a:o an OK implementation?
16:57 rbuels does it need fixing for anything?
16:57 pyrimidine I'm not sure.
16:59 pyrimidine Looks fine to me just glancing through
17:00 pyrimidine There are a few deprecated methods in there from the Feature/Annotation rollback a year ago
17:00 ende joined #bioperl
17:01 * ende waves
17:01 pyrimidine o/
17:01 pyrimidine lunch &
17:14 perl_splut ok, looks like I might be able to automate SpectrumMill :)
17:14 * pyrimidine wonders if that is a type of breakfast cereal
17:14 pyrimidine ;)
17:14 perl_splut Mass Spec analysis softare
17:14 perl_splut software
17:14 pyrimidine ah
17:15 perl_splut takes far too many clicks to go from raw data to output and I don't want to deal with it manually if I can use Mechanize to automate it
17:25 rbuels pyrimidine: what's the approx svn rev number range for the rollback you referred to.
17:25 * pyrimidine looking...
17:26 pyrimidine any reason?
17:30 pyrimidine rbuels: initial commit was 11669, with several more for bugs not caught out by the test suite
17:30 rbuels pyrimidine: just trying to get a feel for what's been going on with the features and such
17:30 rbuels pyrimidine: at sgn, we've been keeping a vendor drop of bp
17:31 rbuels pyrimidine: and it's drifted out of line
17:31 rbuels pyrimidine: had our own patches to bsfa and such.  a bad idea in retrospect
17:31 pyrimidine The problem that rollback dealt with was a lot of core changes added literally just prior to 1.5.0
17:31 pyrimidine (including FeatureIO, BSFA, etc)
17:32 pyrimidine that were committed to main trunk directly and not tested on a branch
17:32 pyrimidine (and w/o consulting the other core devs)
17:32 rbuels was the test suite more spotty then also?
17:32 pyrimidine yes
17:32 rbuels make sense
17:32 pyrimidine (and it arguably still in in a few areas)
19:44 perl_splut kewl. HTML::Form has an undocumented method that can be used to inject new input into a form.
19:44 perl_splut makes it easier to modify a form when one can't use the JS to make the changes
20:54 faceface left #bioperl
23:30 rbuels joined #bioperl

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