blog

Continuing my previous post, much like the “foundation” meeting at the Spring 2005 ACS National Meeting, many of us met over food and drinks to talk about chemistry, informatics, project development, and other things.

There were lots of discussions and I tried to jot down bits. I’m not sure I got everything, but have tried to include the gist of the conversation for those who weren’t there. I do not promise that these will be listed in any sort of order.

• Ajay, from the National Institutes of Health, recently posted to the chemistry informatics list. He was also at a question-and-answer session at the ACS and made it clear that the NIH is very interested in open source chemistry developments beyond PubChem. Concerns were raised at dinner whether such money will go towards new efforts or support for already useful tools (e.g., in the Blue Obelisk).
• Warren discussed how he has been fairly successful marketing PyMol to pharma companies, particularly for very targeted projects or features. So companies are quite comfortable with financial support to open source projects, but more so when there is a clear definition for the money.
• Several people suggested the idea of setting up a US-based non-profit foundation to support Blue Obelisk. This seems like a good idea, although there are some legal details (not so bad) and IRS/accounting details (unclear).
• Warren suggested that for the next ACS meeting (or generally for ACS meetings in the future), that there be a Blue Obelisk booth at the exhibition, showcasing open source chemistry.
• This brought up the general subject of clear demonstrations and Live CDs. I know there is already something of the latter, but demo websites and obvious user-friendly demos would be *excellent* projects for future Blue Obelisk efforts. (e.g., right now we have many toolkits, but perhaps fewer concrete applications, and almost no application integration into a full chemistry workflow).
• Peter discussed itemizing unique strengths in the Blue Obelisk. I have Open Babel (i.e., so many file formats), crystal structures, journal indexing, and Bioclipse listed.
• This lead to the general subject of needs. I mentioned my current effort for a cross-platform desktop 3D editor with the Kalzium project. There’s also a need for an improved 2D editor, e.g., revitalizing/reworking JChemPaint.
• Discussions were made about a need for “professional” software development, including strong regression testing, larger datasets (e.g., Open Babel’s SMILES parser is used for eMolecules and survives almost 5 million compounds).
• ZINC can obviously play a major role in dataset testing since it provides millions of compounds for free use.
• Our guests from pharma admitted that:
  • They’re fed up with commercial chemistry software vendors (lock-in, expensive prices)
  • They lack influence over commercial vendors (!)
  • They feel they have no alternative at the moment
• One continuing discussion/debate was for a funding model from pharma, academia, grant organizations, etc. Current news from the NIH and NSF in the US is encouraging, but we still need to provide a clear benefit in 3-5 years. What is the return on investment for these groups?
• On the same topic, what is the Blue Obelisk “roadmap” for the next 3-5 years? What could we do with investment? We have useful products NOW, but where are all of our projects going with further development and $$$ and manpower.