Depth-First: Encapsulated PostScript for Cheminformatics

Encapsulated PostScript for Cheminformatics

Rich Apodaca — Thu, 07 Aug 2008 10:46:00 +0000

Previous articles have discussed ways to display chemical structures in a variety of vector graphics formats, including Scalable Vector Graphics (SVG), Vector Markup Language (VML), and Shockwave Flash (SWF, or "Flash"). There is one other vector graphics format still in common use that has so far not been discussed: Encapsulated PostScript (EPS).

EPS is a vector graphics format developed by Adobe Systems specifically for desktop publishing. Apparently, the last revision to the EPS Specification, Version 3.0, was published in 1992. Although ancient as technologies go, EPS can still be found in use today. Like the other vector graphics formats discussed here, EPS has all of the features necessary to create high-quality 2D images of chemical structures.

To demonstrate, a development version of ChemWriter was used to convert a molfile representing rosuvastatin (Crestor) into the corresponding color EPS file. The result can be downloaded here.

An open source library, jlibeps, can be used to encode EPS in Java. The business end of the library is an EPS encoding class that extends Graphics2D, enabling most Java2D rendering to be converted into EPS. If you've worked with the Batik SVG toolkit, the concept is similar.

Although jlibeps is a good solution, EPS shares enough similarities to other vector graphics languages that directly encoding EPS output is also a viable option, and the one that was used in the rosuvastatin example above.

Image Credit: Kelly Smith

"Encapsulated PostScript for Cheminformatics" by Rich Apodaca

Tue, 12 Aug 2008 14:10:10 +0000

Gilbert,

Thanks for bringing up mol2ps. I remember looking at it in the context of free chemical database components. I believe I even tried to compile it to Java bytecode using NestedVM and the Pascal compiler at one point, but without success.

I've depicted the rosuvistatin molfile using mol2ps and put the output here for comparison. Looks like PostScript 2.0 output. Although I can open it with gv, it doesn't seem to open with Gimp like the output from ChemWriter does.

"Encapsulated PostScript for Cheminformatics" by gilbert.r.mueller atwebdotde

Tue, 12 Aug 2008 11:03:09 +0000

Prof N. Haider (university of vienna) has published a very nice solution for 2D depiction of molecular structures in ps format, using Ghostscript to convert ps-files to bitmaps: mol2ps It is configurable, licensed under GNU, available for different systems (linux, FreeBSD, Win) and can be googled via "mol2ps haider".

"Encapsulated PostScript for Cheminformatics" by Rajarshi Guha

Mon, 11 Aug 2008 20:52:17 +0000

I actually don't know how one would embed EPS into PDF - the iText library directly generates PDF. I used it in the draw2d example program (http://tinyurl.com/6j2cr4)

"Encapsulated PostScript for Cheminformatics" by Rich Apodaca

Mon, 11 Aug 2008 16:08:49 +0000

Rajarshi, interesting idea. The example does set a BoundingBox of 0 0 300 300.

But I haven' tried to embed these EPSs into PDFs. What else would be necessary?

"Encapsulated PostScript for Cheminformatics" by Rajarshi Guha

Mon, 11 Aug 2008 13:32:49 +0000

It should be relatively simple to provide PDF output as well (using somethinglike iText). One thing that would be very useful is to set the appropriate bounding box, so that rather than having an image on a whole page, the output size would simply be the size of the minimum bounding rectangle.