The idea of open source drug development gained some attention with an article in The Economist back last June. My ISP’s server and backup crash, coupled with my access and backup issues, made that original entry vanish into the ether. However PLOS Medicine (the open source, open access medical journal) has an ongoing series on initiatives in drug discovery for neglected diseases (in this instance, tropical and third world diseases); in its December issue, Stephen Maurer, Arti Rai, and Andrej Sali elaborate on their proposal to use an open source model for the development of drugs that are potentially of great benefit but are not of significant commercial interest for pharmaceutical companies operating on the standard commercial model. They propose using bioinformatics methods to identify possible drug targets, given our knowledge of various pathogenic genomes, and then tap academic and pharmaceutical companies for in-pipeline drugs or volunteer scientific expertise to help develop drugs against those targets.

The original discussion over on Slashdot, on which I first commented, veered off into discussion of intellectual property rights, not surprising given the demographics of the community, their area of expertise (not drug development), and their politics. Setting aside matters of intellectual property, the first part of Maurer et al’s proposal, to use bioinformatics methods to identify drug targets, is plausible: publically accessible bioinformatics resources are extensive and accessible by anyone with a computer, a project, and enough knowledge to be getting on with, and bioinformatics is already being heavily used in the identification of drug targets. The discipline, like most computing disciplines, is evolving so swiftly that the ability to decode indifferent documentation and understand the principles is almost more valuable than knowledge of any particular tool, beyond the core tools. But I do wonder whether life sciences has the demographic to make a meaningful go of this approach: bioinformatics is a young discipline, certainly younger that software engineering, and therefore the pool of practitioners is not as extensive; life sciences are less amenable to being practiced as a hobby and early, and I suspect life scientists would tend to be older and more settled into institutions before they gained sufficient knowledge to make an effective contribution, especially at the highly regulated later stages of drug development. In addition, successive long-term contractions in the life sciences job market have meant that many people who might have made a contribution have left the field over the years. Those are, however, merely impressions; in the time honoured phrase of the discipline, more research is required.

There is no getting away from the fact that the non-clinical and clinical testing phases for drugs are lengthy, expensive and highly regulated, although regulators and others take a special interest in neglected diseases and orphan drugs. Money, expertise and coordination will be required at this stage. Maurer et al mention the idea of “virtual pharma”, such as the Drugs for Neglected Diseases Initiative, whose mandate is to push promising drugs on through the pipeline; the work of the DNDi was described in the first issue of PLOS Medicine. Should the drug developers wish to take the drugs all the way through a regulatory application (which will be necessary if this is a new drug never marketed before), then the writing of a regulatory submission could be done in an open source manner, on-line, using open-source collaborative writing software, wiki, blog, database, or possibly even custom-built for the purpose.

In a letter in this month’s (February’s) issue, Richard Stallman of the free software foundation writes in support of the concept, reminding the reader that the open source software movement has 20 years’ experience in resisting commercial and proprietary interests. Moreover, he points out the possible effect of the open access movement in scholarly and biomedical publishing in advancing such work.