YouTube Facebook LinkedIn Google+ Twitter Xinginstagram rss  

February 10, 2003 | It's a brand-new company, with brand-new technologies and a totally new way of doing things. Infinity Pharmaceuticals Inc.'s CIO, Andy Palmer, couldn't have asked for a cleaner slate.

Clean slate: Building from the ground up, Infinity CIO Andy Palmer is helping the company avoid the silo effect and maximize efficiency.
Instead of stepping into a morass of legacy systems and fiefdoms, Palmer has been able to create a next-generation systems architecture in an "open, knowledge-sharing culture," he says. From e-notebooks, to grids, to knowledge management, Infinity plays with all the innovative tools that other companies have to ponder over because they would require the additional high costs associated with change.

Infinity's systems requirements are fairly unique. "The information systems required to support diversity-oriented synthesis differ from those needed for combinatorial chemistry," Palmer says. "The information we need to gather is more complex, and there is more of it."

In addition to storing information about the basic structure of compounds, Infinity captures all of its experimental results, including the synthetic reactions, purity of the product, and experimental conditions. It also does a lot of molecular modeling. "Creating and using virtual libraries is a critical part of our process," Palmer says. For that they collect a host of 2-D and 3-D properties, such as molecular weight, the number of rotatable bonds, and so on.

Emulating the ICCB model, Infinity expects chemists and biologists to work together. For that to happen, Palmer must "link to all the information on the biology side that is associated with that compound," while measuring the different needs of the two sides. "To chemists, everything is a reaction, and reaction-based templates are logical for them," he says. "Biologists, meanwhile, deal with unstructured information, and the templates have to be much more flexible and open."

To bring everyone to the same plane, Palmer has connected many different systems using a standard vocabulary and a Web services architecture implemented in Java (see "Divide and Distribute," Nov. 2002 Bio·IT World). These Web services are exposed to end users through "dashboards" developed using Microsoft .NET and delivered through a standard portal framework. Homogenous hardware made all of this easier to implement: Besides one Linux cluster from IBM for docking small molecules into proteins, Infinity's desktops and servers are all Dell, and they are all running Microsoft Windows. "Even our large Oracle databases are also running on Windows, and we have seen no need to migrate those," says Palmer, who runs dozens of applications and databases.

That environment should facilitate grid computing. Palmer plans to install a screen saver on staff desktops that continuously docks small molecules into proteins.

The company has also embraced knowledge management and has taken the idea as far as trying to "define the process we call 'chemists' intuition,'" Infinity's applications service manager Andy Ellicott told a Marcus Evans pharmaceutical knowledge management conference. By breaking down researchers' decision-making processes, the company hopes to codify successful practices and avoid the redundancy that frequently plagues research.

Despite the latest IT toys, Palmer knows they won't necessarily yield more drugs: "The things that make a company great haven't changed and don't change."

—Malorye Branca

Back to Conquering Infinity with Chemical Genetics 


For reprints and/or copyright permission, please contact Angela Parsons, 781.972.5467.