Camelia, the Perl 6 bug

IRC log for #cdk, 2009-03-25

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

All times shown according to UTC.

Time Nick Message
00:48 GUERRILLARAWFLAR joined #cdk
00:48 GUERRILLARAWFLAR hey
00:50 GUERRILLARAWFLAR joined #cdk
00:52 GUERRILLARAWFLAR left #cdk
05:25 jbrefort joined #cdk
05:30 egonw joined #cdk
07:36 Gpox joined #cdk
08:43 egonw joined #cdk
08:56 bag joined #cdk
09:57 jbrefort joined #cdk
13:19 s-wolf joined #cdk
13:22 s-wolf Hello, I have a problem regarding the InChI generation in CDK 1.2. When I generate an InChI from a previously read in Mol file the InChI is different from the pubchem one.
13:23 s-wolf I thought the InChI are unique or did I get something wrong here?
13:24 s-wolf for example this InChI generated from the CDK is not read in again: InChI=1/C15H12O5/c16-9-3-1-8(2-4-9)13-7-12(19)15​-11(18)5-10(17)6-14(15)20-13/h1-6,13,16-18H,7H2
13:24 jbrefort InChI spec changed
13:25 jbrefort and PubChem upgraded to the new version
13:27 s-wolf ah ok. But why isnt the INChI reader working on the previously generated one?
13:27 jbrefort I'm not a cdk expert
13:32 s-wolf ok thanks! :)
14:14 egonw hi s-wolf
14:15 egonw s-wolf: which PubChem CID, and where did you get the mol file?
14:16 egonw pubchem is indeed now listing the Standard InChI...
14:16 egonw so, that could be the problem, but I need a bit more info to pinpoint the underlying cause
14:16 egonw I note that the inchi you gave does not have stereo info...
14:17 egonw which is possible... but from the chem formula I'd say there is likely a stereo center in the mol
14:17 s-wolf hi egonw
14:18 s-wolf i got the molfile from kegg (http://www.genome.jp/dbget​-bin/www_bget?cpd+c00509)
14:18 zarah s-wolf's link is also http://tinyurl.com/dh26lr
14:18 s-wolf i read in this molfile and added the hydrogen to it
14:20 s-wolf from this atomcontainer I created the InChI resulting in the string above
14:21 s-wolf InChI warning: Omitted undefined stereo
14:21 s-wolf this warning was printed as well
14:22 egonw it seems that the InchI code in the CDK did not pick up the wedge bond info...
14:23 egonw the MDL reader would have read that
14:23 egonw the rest of the InChI seems to matche quite well actually...
14:24 egonw you mention that the InChI is not read in again...
14:24 egonw into the CDK?
14:25 s-wolf yes....I tried to read that inchi again...the chemSequenceCount is 0
14:26 egonw will have to look at that...
14:26 s-wolf InputStream is = new ByteArrayInputStream("InChI=1S/C15H12O5/c16-9-3-1-​8(2-4-9)13-7-12(19)15-11(18)5-10(17)6-14(15)20-13/​h1-6,13,16-18H,7H2/t13-/m0/s1".getBytes("UTF-8"));
14:26 s-wolf INChIPlainTextReader reader1 = new INChIPlainTextReader(is);
14:26 s-wolf ChemFile chemFile1 = (ChemFile)reader1.read((ChemObject)new ChemFile());
14:26 egonw maybe the reason is that CDK does not support standard inchi again...
14:26 egonw can you try removing the t, m and s layer
14:26 s-wolf sure
14:26 egonw and make is InChI=1/C15...
14:27 s-wolf ahh wrong...thats the generated one: InChI=1/C15H12O5/c16-9-3-1-8(2-4-9)13-7-12(19)15​-11(18)5-10(17)6-14(15)20-13/h1-6,13,16-18H,7H2
14:27 egonw yes :)
14:27 s-wolf the previous one works
14:27 egonw which one does?
14:28 egonw the InChI=1/C15 one?
14:28 s-wolf the one with the InChI=1S
14:28 egonw ah... ok
14:28 egonw interesting...
14:28 s-wolf this is from pubchem....a similar structure
14:28 egonw do you have a SourceForge account? perhaps?
14:28 s-wolf yes
14:28 egonw if, so, please file a bug report
14:29 egonw and I look into it after this week
14:29 s-wolf ok i will do right away... thanks for your help
15:21 s-wolf mhm maybe I found another thing... I just saw when I write "INChI=...." I get a result...but when I write "InChI=..." nothing gets read in....but the previous problem is still there... I will file a bug report
15:32 egonw please add that as note too
15:42 s-wolf ok done
15:44 egonw thanx!
19:15 scoott left #cdk
19:47 jpansanel moin
19:47 zarah privet jpansanel

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