Credit: DE NOVO PHARMACEUTICALS

Most companies seek hits against members of the main families of proteins known to be likely drug targets, such as G-protein-coupled receptors, kinases and proteases. Not surprisingly, achieving selectivity for just one member of such a family can be a considerable challenge.

“Just about anybody can get a hit against a kinase,” says Richard Scott, head of chemoinformatics at De Novo Pharmaceuticals, a company based in Cambridge, UK, that uses virtual screening to discover drug leads. “It's easy to get a hit, but not nearly so easy to get a selective hit.”

As a result, many companies are striving to get to know their target protein families better. This approach has always been central to the philosophy of Vertex Pharmaceuticals, a drug-discovery firm in Cambridge, Massachusetts, but Mark Namchuk, head of high-throughput screening at the company, says that this is now a trend throughout the industry. At Vertex, early discovery is focused on whole protein families rather than individual target proteins.

Emphasizing just how central an understanding of the biology of the target family is to the usefulness of the screening data, the high-throughput screening facility at Vertex is run as a division of the enzymology group, which refers to screening as “high-throughput enzymology”, and to screening campaigns as “experiments”. As well as giving valuable insights into selectivity, “this approach allows you to view today's data as foundations for future projects”, Namchuk says.

Although not yet a replacement for bioassaying the activity of your molecule on the proteins themselves, virtual screening of similar targetrs can help in identifying problems that might crop up further down the pipeline. “Virtual screening can give you a heads-up to other potential interactions,” says Scott Kahn, a senior vice-president at software manufacturer Accelrys in San Diego, California.