Camelia, the Perl 6 bug

IRC log for #cdk, 2011-02-23

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

All times shown according to UTC.

Time Nick Message
02:09 CIA-48 cdk: Egon Willighagen master * r688da8b / (3 files in 3 dirs):
02:09 CIA-48 cdk: Added a signature-based fingerprinter
02:09 CIA-48 cdk: Signed-off-by: Rajarshi Guha <rajarshi.guha@gmail.com> - http://bit.ly/e6x2uT
02:15 CIA-48 cdk: Egon Willighagen master * r37e81e8 / (9 files):
02:15 CIA-48 cdk: Split out unit tests only for fixed-length fingerprints
02:15 CIA-48 cdk: Signed-off-by: Rajarshi Guha <rajarshi.guha@gmail.com> - http://bit.ly/fLhBT4
05:03 egonw joined #cdk
05:36 CIA-48 cdk: Egon Willighagen master * r6fc4803 / src/META-INF/test-signature.cdkdepends : Added missing dependency - http://bit.ly/f5QDNp
05:46 egonw left #cdk
06:15 egonw joined #cdk
06:21 egonw_ joined #cdk
06:21 egonw left #cdk
06:23 egonw__ joined #cdk
06:27 egonw_ left #cdk
06:31 egonw joined #cdk
06:35 egonw__ left #cdk
06:38 egonw_ joined #cdk
06:42 egonw left #cdk
06:46 egonw__ joined #cdk
06:50 egonw_ left #cdk
07:18 egonw joined #cdk
07:22 egonw__ left #cdk
07:24 CIA-48 cdk: Egon Willighagen master * r416a4cc / src/META-INF/test-signature.cdkdepends : Added another missing dependency - http://bit.ly/eORZt8
07:39 egonw_ joined #cdk
07:42 egonw left #cdk
07:46 egonw__ joined #cdk
07:49 egonw_ left #cdk
07:53 sneumann_ joined #cdk
07:58 egonw joined #cdk
08:00 Gpox joined #cdk
08:02 egonw_ joined #cdk
08:03 egonw__ left #cdk
08:03 egonw__ joined #cdk
08:06 egonw left #cdk
08:07 egonw_ left #cdk
08:07 egonw__ left #cdk
08:08 egonw joined #cdk
08:15 egonw left #cdk
08:24 sneumann_ left #cdk
08:36 sneumann_ joined #cdk
08:37 egonw joined #cdk
09:12 stain_ left #cdk
09:12 stain joined #cdk
10:10 jbrefort joined #cdk
12:06 jbrefort left #cdk
13:50 bag_ left #cdk
14:02 bag_ joined #cdk
14:34 maclean joined #cdk
14:35 maclean hi
14:35 egonw hi maclean
14:35 maclean hah, you beat zarah :)
14:35 egonw zarah: you awake?
14:36 egonw I think she died
14:36 * maclean is splitting an SMSD mega-patch into separate commits.
14:36 maclean We can rebuild her!
14:36 maclean (or, well, maybe you can...)
14:38 egonw cool :)
14:43 maclean Oh, and I saw the signature-fp by the way. Looks fine. I was thinking though, that the greater the signature depth, the more redundant info there is.
14:43 egonw yes, hence the constructor allowing to set it
14:45 maclean sure.
14:46 maclean Perhaps could be mentioned in javadocs? Also, that a height (er..depth)-2 is probably 'good enough' :)
14:46 maclean Although, to be honest, I have no proof of this.
14:47 egonw likely depends on the data set too
14:49 maclean I suppose. I guess there might be weird 'globally non-local' graphs in some chemical sets.
14:53 maclean At one point I considered a sort of 'chemical-space FP'  made from frequencies of signatures counted across all the molecules in the set. Not sure if it would be useful though.
15:00 egonw maclean: like this? http://chem-bla-ics.blogspot.com/2010/​06/qspr-modeling-with-signatures.html
15:01 maclean I was about to say "I'm sure I read that.." when I saw my comment.
15:02 maclean But, anyway : yes and no. The signature 'vector' ( like [70:1, 54:1, 23:1, 22:1, 9:9, 45:2]) is made per-molecule
15:02 maclean Rather than for all the molecules.
15:03 maclean I'm talking about comparing two sets of molecules, not the molecules in a set.
15:04 egonw the bit positions are unique for all molecules
15:04 egonw oh, like that...
15:04 egonw sorry
15:04 egonw yeah, easily publishable, I think...
15:04 egonw comparing data sets is still what people publish
15:06 maclean I would have thought there should be a scientific question, like "Are these drugs different from natural products?" or some such.
15:07 egonw the business model is basically, does this data set bring in new chemistry? if so, price *= 5
15:09 maclean Heh.
15:11 maclean Oh, and it would be quite easy to make a SignatureQuotientGraphFingerprint too. Again, not sure how meaningful it would be, but I don't get the impression that is the point of FPs..
15:17 egonw hahahah
15:18 maclean I'm probably being a little unfair...
15:21 jbrefort joined #cdk
15:27 egonw left #cdk
15:38 maclean left #cdk
15:49 egonw joined #cdk
16:41 egonw left #cdk
16:47 sneumann_ left #cdk
20:52 zarah left #cdk
20:52 zarah joined #cdk
20:54 Gpox left #cdk
21:48 bag_ left #cdk
21:48 s_wolf left #cdk
21:48 Conrad left #cdk
21:48 zarah left #cdk
21:48 CIA-48 left #cdk
21:49 CIA-48 joined #cdk
21:56 jbrefort left #cdk
22:37 _ilbot joined #cdk
22:37 Topic for #cdk is now The Chemistry Development Kit (http://cdk.sf.net) | JUnit CDK 1.2.7: 15690, 76 fails + 15 errors (0.58%) | Logs at http://moritz.faui2k3.org/​irclog/out.pl?channel=cdk | CDK News Feed: http://friendfeed.com/rooms​/chemistry-development-kit

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