By Kevin Davies

May 16, 2008 | In an unusual move for big pharma, Eli Lilly last month made its Discovery IT platform – known internally as the Lilly Science Grid (LSG) – open source. The initiative could spark new interest within the biopharma community for sharing pre-competitive content and software.

“This is something that John Reynders was interested in and has been in the works for a while,” says Susie Stephens, principal research scientist with Lilly. (Reynders left Lilly for Johnson & Johnson last year (see Reynders Takes the CIO Reins at J&J R&D, Bio-IT World, Dec. 2007.) The move addresses a key question: “Do we [biopharma] all really need to manage very similar resources or are there areas where we can work together? “ 

Susie Stephens

LSG, which began in 2005, has a core framework to enable users to build plug-ins to create applications. “This is our core discovery IT framework, and was developed on the biology side, but has extended into chemistry internally,” explains Stephens. The current lead technologist is Andy Ring.

“Eighteen months ago, John Reynders thought it would be good if we made LSG open source, which we’ve just done,” says Stephens. The software is available on Sourceforge.net. “It’s all well and good developing plug-ins internally for the framework, but if there are many people within the broader life sciences community doing that, it could make it much more interesting,” she says.

The publicly available version of LSG – dubbed Life Sciences Grid so as not to upset other pharma companies – includes a select group of “less proprietary” plug-ins, including those for Gene Browser, NCBI Entrez, and Gene Ontology. “We’re also kick-starting some collaborations, working with people in academia to develop plug-ins to work in the framework,” says Stephens. Indeed, she says at times Lilly’s open-source ambitions are exceeding those of some potential academic partners. “We’d like to see many people building plug-ins, some of which we wouldn’t have thought to have done.”

But why go to the trouble? “Lilly generally is looking very closely at where it makes sense to have capability in house versus collaborating,” says Stephens. “We’re looking to become much more networked as a company, so we’re looking hard at what we consider proprietary and what we must keep to ourselves, and what are we actually doing that other pharma companies are also doing that we consider to be pre-competitive.”

Adds Stephens: “We don’t think it makes financial sense for all pharma and biotech companies to all be developing a core discovery IT framework, which pretty much does the same thing. If we could all work together using this core framework, then it would be good for everybody.“

Pharma Precedents
“I’m not aware of a huge number of projects where pharma is making code available open source,” says Stephens.  She cites PISTOIA -- a new initiative to share data models, and web-service interfaces in areas such as cheminformatics, which is also available through Sourceforge. But, “It is fairly unusual to be putting a core framework out there.  I’m not aware of pharma companies sharing software.”

Stephens, who formerly was something of a semantic web guru at Oracle, is enjoying her matchmaking role at Lilly. She focuses chiefly on open innovation, identifying areas “where it makes sense for Lilly to establish external collaborations”  and gaps in the company’s in-house capabilities or resources where it would make sense to find collaborators, in areas such as informatics and the semantic web.

Another goal is to open a small “open innovation center,” probably in Cambridge, MA, spun off from Lilly that would focus on pre-competitive collaborations across pharma, again in areas such as software and discovery IT. Stephens is currently researching funding sources.

LSG can be found at www.sourceforge.net (search “LSG”), under a Berkeley Software Development license.

What Is LSG? The Lilly Science Grid is a software infrastructure that Lilly developed internally for drug discovery programs. LSG is a plug-in hosting and deployment framework that sits on top of Microsoft’s Composite Application Block. LSG is a rich client that requires .NET 2.0 or higher. The framework simplifies the task of creating new plug-ins by providing a Visual Studio template from which developers can quickly learn and expand. Users can choose which applications and plug-ins to use within an integrated environment. Within Lilly, Stephens says LSG is mainly used by bioinformaticians and computational biologists, but it has many additional capabilities, including target assessment. She stresses its value as a platform for interoperability. People interested in working with the open-source version can take the code, develop enhancements and modifications, there are restrictions on further commercialization. Before going open source, Lilly carefully parsed the LSG code “to make sure it didn’t contain anyone else’s code,” says Stephens. It used to be dependent on Oracle databases, but now works equally well on MySQL. -- K.D.