> My current conversion: .cdx -> .xyz using make3D. And after this, checking > image from .xyz using chemcraft. And for big structures it looks not enough > good.
It would really help to have actual structures (e.g., both ChemDraw and XYZ) to understand what you mean by "not good enough." (BTW, I'd call your compounds small molecules, so it's a relative term.) > Maybe are there any articles that were used for writing this operation in > open babel? I'm a bit upset that I could't find anyone. > Please, if you know about that, could you send me a link with more detailed > description of 3D conversion? I'd be very grateful. The 3D generation process in Open Babel certainly wasn't a publishable result. It's a rule-based method with a relatively small pool of ring fragments. My rough guess, given the bitmap 2D diagrams is that the rings you have aren't well represented. Typically that's handled with forcefield clean-up, and depending on what you've provided, you may need more cleanup. A Google Summer of Code project increased the size of the fragment database and worked on other elements of the 3D generation code. That will be a publishable result - we're currently working on the draft of the article. That code, naturally, is in the 'master' branch on GitHub. Again - if you can provide actual XYZ or ChemDraw structures and a bit more specifics on "not good enough" it would help considerably. Best regards, -Geoff _______________________________________________ OpenBabel-discuss mailing list OpenBabel-discuss@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/openbabel-discuss
