Atom overlap in depictions

[Dale Cameron]

The depictions coming up from my smiles strings seem to have a large amount of overlapping in the structure images.  I searched online for some public examples that fail in a similar fashion and I found a few.  Here are two.

Example 1: CC1(C)(CC(=O)C(C(=O)C1)C(c2ccc(o2)[N+](=O)[O-])C=3C(=O)CC(C)(C)CC=3(O)) (BAS 03161502 from Asinex collection)

See this link for example of non-overlaped structure on pubchem.  http://pubchem.ncbi.nlm.nih.gov/summary/summary.cgi?cid=3145896

Example 2: O=C2C1=CC=CC=C1[C@H](/[C@](C(C4=CC=CC=C4F)=O)=C(O)\C(NC3=CC(C(C)=O)=CC=C3)=O)O2 (NSC 640982)

See this link for example of non-overlaped structure on pubchem.  http://pubchem.ncbi.nlm.nih.gov/summary/summary.cgi?cid=369251&loc=ec_rcs

The attached file is a jpg with the two depictions from LICCS in them (left is example 2, right is example 1.

Anything I can do about that?

I'm using the most recent version of LICCS with Excel 2010 and Windows 7.

Thanks in advance,

Dale

[Kevin Lawson]

Hi Dale

I am going to sound like a broken record here but, analogously with your previous stereochemistry problem, structure depiction in LICSS is taken care of by JChemPaint (the chemistry editor/depictor of CDK) and the client programmer has little control over the actual layout of depictions.  I checked that 'native' JChemPaint (which you can get at by executing JChemPaint-3.2.0.jar in your c:\services\licss-2.2 directory) also lays out the structures with overlapping atoms and that the 'clean' facility doesn't improve things - same less-than-perfect results I'm afraid.  You might like to look at the cdk-jchempaint mailing list for similar problems and, perhaps, post your examples there?

Sorry not to be able to help

Best wishes

Kevin

[Dale Cameron]

I kind of figured that would be the case but wanted to start here first. Thanks a bunch Kevin.  

Dale

[Dale Cameron]

Just updating.  I did post this problem to cdk and jchempaint bug reports and just this week cdk closed this one stating that the newest code base for cdk handles them correctly.  I agree as I checked online and that is true.  However, the latest version of jchempaint still does not handle these correctly, so this problem persists.  I was then told that jchempaint is no longer in active development so a new version capturing the latest codebase of cdk, which fixes the problem, will likely never materialize.

Is there any plan to move from jchempaint to cdk depict or some otehr format for teh future?

Great tool by the way!

Dale

[John May]

I don't use Excel/Windows so can't check but do you need structure editing? There's is some long history here but basically JChemPaint shouldn't be doing any rendering! this should all be handled by the CDK. You only need JChemPaint for the editing.

I updated JChemPaint a couple of years ago to the latest API but you need to checkout and build the code yourself. I'll update to the very latest SNAPSHOT now so we get the layout improvements but since there isn't a place of Java Applets any more I'm not motivated to do a release.

John

[Kevin Lawson]

Hi Dale

I have just released v3.3 of LICSS which uses the very latest version of the CDK for structure depiction and JChemPaint only for structure editing. Many thanks to John May who did the necessary recoding.

Let me know how you get on.

Best wishes

Kevin