The genomics revolution has provided innumerable potential targets for new drugs. But many of these disease-causing proteins seem to be "undruggable": companies can't find compounds that inhibit them. Into this breach comes Boston startup Infinity Pharmaceuticals, which plans to use new chemical-synthesis methods to build more varied and complex drug candidates than have previously been possible.

Infinity's technique, pioneered by Harvard University chemist and company cofounder Stuart Schreiber, involves binding simple starting compounds to hosts of small plastic supports. The compounds acquire new chemical groups as they are exposed to a series of carefully chosen solutions; the supports are separated at various steps, producing smaller collections bearing unique chemicals. Such reactions are easier to track than those performed in solution and can yield a wide variety of products.