Camelia, the Perl 6 bug

IRC log for #bioperl, 2009-05-13

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

All times shown according to UTC.

Time Nick Message
03:54 rbuels joined #bioperl
09:50 emilbarton joined #bioperl
14:57 rbuels joined #bioperl
15:48 faceface yo rbuels
15:48 faceface DBTAG:ID". The DBTAG indicates which database the referenced object can be found in, and ID indicates the identifier of the object within that database
15:48 faceface what is the right DBTAG for a genbank accession?
15:48 rbuels erm.  i don't know if there is a standard one
15:49 faceface oh
15:49 * faceface cries
15:49 rbuels welcome to bioinformatics.
15:49 faceface ;-)
15:49 * faceface sees NCBI_GI... and goes for NCBI_ACCN
15:49 faceface here btw http://www.sequenceontology.org/gff3.shtml
15:50 faceface figure if I'm doing it, I may as well try to make it right...
15:50 faceface also... how can I combine matches to the 'contigs' with a PGP which orders said contigs?
15:50 faceface in a gff ...
15:51 faceface me also defines your BACs thus... http://www.sequenceontology.org/wiki/i​ndex.php/Category:SO:0000149_!_contig
15:51 rbuels what's the 'P' in PGP?
15:51 faceface er... PrettyGood?
15:51 perl_splut yep
15:51 rbuels oh.
15:51 faceface Pretend?
15:51 perl_splut PrettyGoodPrivacy
15:51 faceface Pray
15:51 faceface perl_splut: its pseudo golden path
15:51 rbuels yeah, what makes is pseudo
15:52 rbuels it pseudo i mean
15:53 rbuels couldn't you just call the BACs BACs?
15:53 rbuels i think that's in SOFA
15:55 faceface C00HBa0012M01.2 contig  SGN     1       118516  .       .       .       Dbxref "NCBI_accn:CU915724.2"  ; ID "C00HBa0012M01.2"  ; Name "C00HBa0012M01.2"  ; Ontology_term "SO:0000149"
15:55 perl_splut Secher Nbiw
15:55 faceface I don't find it in sofa
15:55 faceface perl_splut: I beg your pardon?
15:56 faceface rbuels: http://www.geneontology.or​g/cgi-bin/xrefs.cgi#ref-G <-- you need to join #bioinformatics ;-)
15:56 rbuels where is #bioinformatics
15:57 perl_splut Secher Nbiw: The Golden Path
15:57 faceface rbuels: #bioinformatics
15:58 rbuels oh.  just assumed it was not here because your whois doesn't say you are in it
15:58 faceface oh... I'm there
16:15 rbuels $featureobj->gff_string, that is an old thing no?
16:16 rbuels (if anybody but me and faceface are listening)
16:16 faceface I'm listening
16:17 faceface I wonder if I need to dump it via a gff object of some kind
16:17 rbuels that's the way i do it
16:17 rbuels FeatureIO
16:17 faceface ahh
16:17 faceface I have a hard time navigating the modules...
16:18 perl_splut use the deobfuscator
16:19 rbuels my $gff3_out = Bio::FeatureIO->new( -type => 'gff', -version => 3) or somesuch
16:19 rbuels at least that's what it used to be
16:24 faceface MSG: only Bio::SeqFeature::Annotated objects are writeable
16:24 faceface rbuels: right, that is what I just got to
16:25 faceface my $gffIO =
16:25 faceface new Bio::FeatureIO( -format => 'GFF',
16:25 faceface -version => 3
16:25 faceface );
16:25 rbuels yarp
16:25 faceface print $gffIO->write_feature($f);
16:25 rbuels that's a little bit of an opaque message
16:25 rbuels i guess Bio::SeqFeature::Annotated are still the strongly-typed features
16:25 faceface Use of uninitialized value in pattern match (m//) at /local/Scratch/dbolser/perl5/l​ib/perl5/Bio/FeatureIO/gff.pm line 107, <FB> line 688.
16:25 faceface ------------- EXCEPTION: Bio::Root::Exception -------------
16:25 faceface MSG: only Bio::SeqFeature::Annotated objects are writeable
16:25 faceface STACK: Error::throw
16:25 rbuels paste it
16:25 faceface STACK: Bio::Root::Root::throw /local/Scratch/dbolser/perl5/l​ib/perl5/Bio/Root/Root.pm:368
16:25 rbuels fewl
16:25 faceface STACK: Bio::FeatureIO::gff::write_feature /local/Scratch/dbolser/perl5/lib​/perl5/Bio/FeatureIO/gff.pm:272
16:25 rbuels above
16:25 faceface STACK: ./fasta2gff_landmarks.plx:72
16:26 faceface ------------------------------​-----------------------------
16:26 rbuels http://sial.org/pbot/
16:26 faceface I figure there is such low traffic in here...
16:26 faceface anyone mind?
16:26 rbuels well, you're probably right about that
16:26 rbuels heh
16:26 * faceface weeps quietly
16:26 * deafferret screams bloody murder
16:26 perl_splut might be low traffic, but pastebots make it easier to read
16:26 faceface hey deafferret
16:26 * rbuels waves
16:27 * deafferret is too stunned to respond
16:27 faceface okey, duly noted
16:27 deafferret paste ... burning my eyes!!
16:27 * deafferret wails
16:27 rbuels hahahaha
16:27 * faceface accidentally deletes all mental notes
16:27 faceface right... I feel another post to bioperl coming on...
16:28 rbuels i was kind of wondering, actually, what the status is of the strongly-typed-features debate
16:29 faceface rbuels: http://www.bioperl.org/wiki/GFF_code_audit
16:29 rbuels cause that's what this is
16:29 faceface rbuels: http://www.bioperl.org/wiki/GFF_Refactor
16:29 rbuels Bio::SeqFeature::Annotated is/was a strong typed feature
16:29 rbuels meaning its feature type is constrained to be in SOFA or whatever ont.
16:29 rbuels you can/could set the ontology i beliece
16:30 faceface rbuels: I just want to print the gff...
16:31 rbuels well if you want to use FeatureIO you have to have BSFA features i believe
16:31 rbuels see, everything i know about bioperl is about a year old
16:32 rbuels so i feel like i shouldn't really be expounding about this because my knowledge is probably wrong now
16:32 rbuels since i have not kept myself plugged in
16:32 rbuels (but i'm working on doing a better job of that)
16:33 faceface hehe
16:33 faceface everything I know about bioperl is nothing
16:36 rbuels i feel another list post from you coming on
16:36 rbuels or an archive search and feeding me links
16:36 rbuels hehe
16:36 rbuels cause i want to know too
16:37 faceface lets see... just been invited to the pub, so I'm in a post and pissoff mood ;-)
16:37 rbuels hehee
16:37 rbuels ok do that then
16:37 rbuels Pdx.pm meeting is tonight so i get to go the pub too
16:37 rbuels it's only 9:30 in the morning here right now
16:37 faceface ace
16:38 perl_splut 4 more hours till early release from work due to Obama
16:39 deafferret perl_splut: wut?
16:43 perl_splut he's speaking here, and because of the traffic nightmare that is going to cause, they have given approval for all employees to leave early
16:43 deafferret ah. i c
16:43 faceface he's such a great guy!
16:43 perl_splut yeah, sure, whatever
17:02 faceface perl -e "use Republican; vote()... humor... exit(1)"
17:03 rbuels is humor() an abstract method of Republican?
17:03 rbuels i don't think it's been implemented
17:04 faceface there it is!
17:04 * rbuels *bonk*
17:04 * faceface has no geek humor skills
17:04 rbuels faceface: aren't you supposed to be off to the pub?
17:04 faceface tes
17:04 faceface yes
17:04 faceface it taks me soo long to write emails
17:06 faceface recently I've been reading this...
17:06 faceface dam... slow web...
17:06 faceface http://producingoss.com/
17:09 faceface Producing  Open Source Software
17:09 faceface How to Run a Successful Free Software Project
17:09 faceface ++
17:09 faceface tells you how to post to mailing list for example ;-)
17:09 faceface 'you are what you write' <- this section got me scared!
17:09 faceface faceface: "is an annoying twit"
17:09 faceface ;-)
17:09 rbuels har har
17:10 rbuels rbuels: mouths off at everybody and doesn't produce any useful code
17:17 rbuels actually, i do produce useful code it's just all within sgn
17:18 rbuels but i'm working on that too
17:23 faceface rbuels: make the transition ;-)
17:23 faceface I'm off for real now - have a nice day
17:24 rbuels nighty night
17:44 deafferret rbuels: the trick is trying to genericize some of that goodness out to CPAN
17:44 rbuels deafferret: that's assuming that is IS goodness
17:44 rbuels hehe
17:44 deafferret / bioperl
17:44 rbuels which it mostly ain't
17:45 deafferret ya, but maybe you have a few neato nuggets
17:45 rbuels yeah we do i'm just being a boob
17:45 deafferret the next time you touch those you could accidentally CPAN them
17:45 deafferret mmm... boob
17:46 rbuels yeah, i'm trying to figure out a way to construct our development environment such that we can distribute our own stuff in a sane way
17:46 rbuels and also have a well-defined path for migrating changes upstream to projects we use
17:46 rbuels like gbrowse and bioperl and whatnot
17:47 deafferret ya. I'm using github for my "oh, perhaps $external_value > 0" stuff
17:47 deafferret and the good ones end up on CPAN
17:48 deafferret (we have one huge-mongous "Omni2" SVN checkout point internally)
20:08 rbuels deafferret: what kind of stuff do you tend to do/write at your bioinformatics gig at unomaha
20:17 deafferret uh, whatever they want that day
20:17 deafferret SearchIO stuff, SeqIO stuff, custom regexy subseq stuff
20:17 perl_splut standard bioinfo stuff, heheh
20:18 deafferret https://rtg.ist.unomaha.edu/ guest guest
20:18 deafferret ya. nothing terribly exciting. researchers here have a hard-on for BLAST alternatives, which bores me
20:18 deafferret "alignment free" :(
20:19 perl_splut heheh
20:19 rbuels expound on this BLAST alternatives business
20:19 rbuels if you would kind sir
20:19 deafferret uh
20:19 perl_splut most of the researchers here don't really know much about the tools available
20:20 rbuels deafferret: i mean examples or something
20:22 perl_splut heheh... I can see that it takes longer and longer to add the next successive sequence to an expanding SeqFeature::Generic object... wonder how slow it will be when it has all 45,000 sequences stored in it :)
20:23 rbuels you probably will lose patience before you get there
20:24 perl_splut yep, I did
20:24 perl_splut so I changed a bit of my code to spare me
20:25 deafferret PubMed ID 15119348
20:25 perl_splut faster to hold the sequence as a string outside the object until done adding all the sequences together :)
20:25 * rbuels is the Swami of Slowness
20:26 perl_splut not sure splitting this large fasta file up into smaller files for each sequence would be any better... 45,000 files in a folder hurts... a lot
20:29 rbuels uh oh, he's gonna reach for making a balanced tree of dirs, or a relational db
20:29 rbuels or AnyDBM actually maybe
20:29 rbuels that might be ok
20:29 perl_splut not in the next 30 minutes
20:38 perl_splut still slows down, but not nearly as much
20:39 perl_splut memory usage still isn't too bad... only about 150MB
20:58 perl_splut wow... 30 minutes in and it's still reading the input file... hasn't even moved on yet
21:04 rbuels one of the great tensions of bioinformatics
21:04 rbuels performance versus nice high-level code
21:05 rbuels well, that gets to be a tension as soon as the person programming knows there is such a thing as nice high-level code

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