After some questions from Jean-Claude, I worked a bit on the CMLRSS feed. It now links to a page with blog items for that specific molecule. When doing that, I also noted that it was not listing the right molecules, so fixed that too; it now again shows the molecules most recently blogged about.
I finally found a bit of time to further work on the molecules part of the postgenomic.com software, and replaced the 'hot tags' tab on the front page, by the eight latest blogged about molecules. Additionally, the system now allows deep linking to molecules in the database via the InChI, for example for N,N,N',N'-tetramethylnaphthalene-1,8-diamine [1]. This is what the front page now looks like:
1.InChI=1/C14H18N2/c1-15(2)12-9-5-7-11-8-6-10-13(14(11)12)16(3)4/h5-10H,1-4H3
Journal titles are extracted by the postgenomic.com software from CrossRef by default. However, ACS journals are not listed there, resulting in a situation that article titles are not given. Now, based on an example script from Euan, I set up a handle_acs.pl script to extract article titles differently. Over the weeks, however, more and more 'Unknown Title's showed up. Today, I added a few more ACS journal titles, and used the script to write equivalents for PubMed (which are not always picked up correctly) and BioMed Central. Most of the unknown titles are now correctly given again.
I noticed last Wednesday that not all blog items are reaching the webpage... there are found and stored in the database, but then nothings happens... But I have not had time to further explore the cause, but will look at this tomorrow again. Sorry for the inconvenience.
Update: I tracked down the problem to a RDF which was no longer a RDF feed, but now showed HTML content that the page was no longer found.
Chemical blogspace is getting more media attention, not the least from C&EN's article Bloggers Anonymous. Here's the list of recently added blogs:
My Chem-bla-ics blog contains a couple of blog items about CMLRSS, that is RSS feeds enriched with Chemical Markup Language. Some time ago, I wrote a plugin for CMLRSS for Bioclipse, as replacement for the Jmol and JChemPaint plugins shown in the CMLRSS article (DOI:10.1021/ci034244p).
Now, since I was looking at some bug reports to fix for the upcoming Bioclipse 1.0.1 release, I discovered that the plugin did not work well with Atom 1.0 feeds from blogger.com. Because Chemical blogspace uses Atom 1.0 too, I wanted to extend the feed with latest molecules with CML content; that is, to make it a nice CMLRSS feed.
Molecules in Chemical blogspace
I mentioned the InChI extension of Chemical blogspace before, and worked a bit more on it today. For example, I only today discovered that the 'pagination' was not working. And, for some days, I wanted the Molecules page to show all molecules being discussed, not only those that are known in PubChem.
Finally, I fixed the sorting, so that the molecules are now sorted based on the post date of the blog item that cites the molecule. It now looks like this (note the molecule count, and that the second InChI does not have an image):
The CMLRSS feed of Chemical blogspace
The next step was to actually include CML in this feed. The CML is created with OpenBabel from a MDL molfile downloaded from PubChem. BTW, this CML can be accessed from the Molecules page too, as can be seen in the above screenshot.
After some tweaking of the Atom feed of Chemical blogspace, and having Bioclipse work with URLConnections for web servers that do not return a Content-Length: field, this was the result:
Yvonne Martin recently wrote about, in her experience, what works and what does not work in chemoinformatics/computational chemistry (DOI:10.1002/qsar.200610102). She based her conclusions on the 17 month usage of services provided via webpages at Abbott. In that period, almost 3 million (unique?) molecules were processed, for which the following properties were requested most, among a few others:
