Camelia, the Perl 6 bug

IRC log for #bioclipse, 2007-09-17

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

All times shown according to UTC.

Time Nick Message
06:04 EskilA joined #bioclipse
07:35 steinbeck joined #bioclipse
07:39 thomas_ku joined #bioclipse
07:55 olas joined #bioclipse
07:56 olas moin
08:09 jonalv joined #bioclipse
08:19 egonw joined #bioclipse
08:20 CIA-31 bioclipse: egonw * r3758 /trunk/plugins/net.bioclipse.data.drugbank/bin/: Removed autogenerated Eclipse dir.
08:24 olas egonw: we are building a fragment library for Bioclipse
08:24 olas fragment is now just a molecule
08:25 olas Since if you download fragment libarries from companies (you want to order your fragments later) they come in just one big SDF
08:28 olas egonw: will it be useful to use SMARTS when doing substructure search or fingerprinter + full substr. ?
08:28 egonw both
08:29 egonw fingerprint is much faster and used as first filter
08:29 egonw then SMARTS for the exact matching...
08:29 egonw FP is not exact...
08:29 egonw that is
08:29 egonw FP returns false positives
08:29 egonw (but does not have false negatives... which is very important :)
08:29 olas ok, so FP + SMARTS is the preferred way?
08:29 egonw yes
08:30 olas not the isomorphism algo?
08:30 olas if you just want to do a substructure search for a small fragment, is SMARTS only enough or should we always use FP first?
08:31 egonw for large libraries you do FP first
08:31 egonw not sure when it starts paying off...
08:31 olas What is large?
08:31 egonw but >1000 definately
08:31 olas ok
08:31 olas probably in most cases then
08:32 egonw would make an interesting CDK News article...
08:32 olas We'll think about it
08:32 olas jonalv will do the DB implementation
08:32 olas I will assist
08:33 jonalv so fingerprinting can be used to identify substructures? The fingerprinting algorithm seems fascinating.
08:33 olas And it will most probably be done with Ibatis as persistence
08:35 jonalv yes and we will probably not use emf :(
08:37 olas no
08:37 olas but this is just a small plugin with a simple DB
08:38 egonw why not reuse the db stuff tobias wrote?
08:39 olas tobias does not support it
08:39 olas and we are getting tired of hibernate
08:40 olas just too many problems
08:40 olas but will definitely look at it
08:40 egonw ack
08:40 egonw :(
08:40 jonalv Yes so we are switching to iBatis. weclone all new problems :)
08:40 jonalv hm welcome (no cloning going on here)
08:40 egonw please email about iBatis on the BC devel list...
08:40 egonw with some links etc
08:41 olas We'll set up a wiki about it
08:41 olas It's really much more reliable
08:41 jonalv wiki or iBatis?
08:41 olas since you get control over your SQL
08:41 olas ibatis
08:41 olas :-)
08:42 olas Jonalv is our hibernate/ibatis master
08:42 jonalv Well Ihave just started learning iBatis but I guess that makes me the master of it in among said group...
09:16 thomas_ku_ joined #bioclipse
10:28 CIA-31 bioclipse: jonalv * r3759 /trunk/playground/plugins/net.bioclipse.keyring/s​rc/net/bioclipse/dialogs/KeyRingLoginDialog.java: Changed so that the login call can be made from a separate non standard swt thread
10:40 CIA-31 bioclipse: jonalv * r3760 /trunk/playground/plugins/net.bioclipse.keyring/s​rc/net/bioclipse/dialogs/KeyRingLoginDialog.java: Cleared out some confusion about sandbox instance and main instance during login / creation of user from the login window.
11:16 EskilA_ joined #bioclipse
11:26 olas left #bioclipse
14:57 EskilA joined #bioclipse
17:10 Nout joined #bioclipse
19:19 edrin joined #bioclipse

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