Camelia, the Perl 6 bug

IRC log for #bioperl, 2009-07-02

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

All times shown according to UTC.

Time Nick Message
12:31 alper_ joined #bioperl
15:54 ende joined #bioperl
16:16 ende anyone familiar with well maintained yum repos for bioinformatics software?
16:22 deafferret not I. "well maintained bioinformatics software" is oft hard to come by at all, let alone yum  :)
16:23 deafferret what the heck is chrM in mouse?
16:23 ende the Mighty chromosome
16:24 deafferret chromosomes 1-19, X, Y: sure. M: ??
16:24 ende found only in Mighty Mouse
16:24 deafferret :)
16:24 deafferret chrMM
17:11 rbuels deafferret: it's probably mitochondria seq
17:15 deafferret ya, you've just reconfirmed that rumor so I'm running with it
17:16 ende mightyocondria?
17:16 deafferret wondering if I can set up a massive search space network with Hadoop on my cluster for this mouse thing... http://hadoop.apache.org/core/
17:39 deafferret now wondering how easy or not my own Erlang farm would be
17:46 ende left #bioperl
18:07 ende joined #bioperl
18:07 ende bioperl-bot translate atgaaaaatcaaattgttgggtataaaaagttcacggttcaaaaaat​ctggaaagcaacaaaagcgcaattagtatttgaaggtgctagaatgg​actacgctagatacctaaaaatgaaaaaacaagggaaagaaatgggt​attacagtaccaaaattcacgttgaaagaaatgtaccaaaatcgtgc​tagcatgaaacgtgatgaactaagaaaagacatgttcctgtgccttg​atattttaaatgaccgagtaggaaaataaaatgtggtttgtaattat​tggtgccagctttggattaattttaggactacgaaagtacaacaagg​gaaaaggagcggaaagagatgagcaaaaaaaatcttagtattagtaa​aatgaaagaactttgtcaaaaagaaggaagtactttctttgaaccta
18:07 bioperl-bot MKNQIVGYKKFTVQKIWKATKAQLVFEGARMDYARYLKMKKQGKEMG​ITVPKFTLKEMYQNRASMKRDELRKDMFLCLDILNDRVGK*NVVCNY​WCQLWINFRTTKVQQGKRSGKR*AKKILVLVK*KNFVKKKEVLSLNL
18:07 ende I actually found a use for that
18:23 deafferret a use for that protein? is it minty fresh?
18:25 ende no a use for the translate function on bioperl-bot
18:26 deafferret :)
18:47 ende wow, ucdavis ftp for seqio (fmtseq) no longer exists
21:31 deafferret I started writing a README  http://github.com/jhannah/​sandbox/tree/master/erlang
22:22 deafferret started the API  http://github.com/jhannah/​sandbox/tree/master/erlang
22:31 rbuels so you have 2M seqs, you want to mutate them according to some rules (generating a 3e3*2e6=6e9 set of possibly duplicated 18-mers), and then remove duplicates?
22:32 rbuels if i'm understanding right, seems like the tack to take would be to parallelize the mutation step, gathering all possible mutations into one big file, then do a sort | uniq on the result file
22:32 rbuels deafferret:  ^^^^^
22:33 deafferret rbuels: ya. harvard pre-prepped big files of all the mutants and then pushed files around on their cluster
22:34 deafferret I (mistakenly?) think that a TCP/IP activated client farm can do the same thing faster w/o pre-prepping
22:34 rbuels the pre-prepping would be the thing to parallelize
22:34 rbuels well, you could parallelize the sort too, actually
22:34 rbuels i think...
22:35 deafferret ya... you have to cause the files you have to sort | uniq | join are way too big to ever work on a single box  :)
22:35 deafferret http://mad.printf.net/MSCC_matc​hing_instructions/matching.html
22:35 rbuels nah.  unix sort can do it
22:35 rbuels it'll use tempfiles, but it can do it
22:35 deafferret "Because the later steps (eg. sorting) will break if the initial set is too large, we work with a small set at a time...."
22:35 deafferret besides, erlang should be fun  :)
22:36 * rbuels does not believe the sort will break
22:36 rbuels well you're right about that
22:36 * deafferret the messenger dodges the bullet
22:37 * rbuels goes and reads about this erlang business
22:37 deafferret http://erlang.org/download/getting_started-5.4.pdf
22:39 rbuels or you could do it in java and use hadoop
22:39 rbuels that should be fun too
22:39 deafferret ya, I stared at hadoop for an hour today and couldn't seem to swallow the java
22:40 * deafferret needs work on his gag reflex
22:40 rbuels at this rate, you'll never be promoted to fluffer.
22:40 deafferret you should join #omaha.dev . stesla is erlang god helping me
22:40 deafferret we're over there chatting about this mess
22:41 * rbuels will lurk
23:48 alper_ joined #bioperl

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