By Kevin Davies

Feb. 1, 2008 | The Canadian developers of a popular drug database have partnered with a software company in an unusual win-win relationship. The deal exchanges short-term funding that might ensure the long-term survival of the database in return for exclusive data access to the commercial partner, GenomeQuest.

DrugBank — a database of drugs, their properties, targets, and metabolizing enzymes — was originally developed in 2005 by David Wishart’s group at the University of Alberta. Wishart’s team also produced the Human Metabolome Database (See The Human Metabolome Project, Bio•IT World, April 2007).

Wishart’s group recently unveiled DrugBank version 2.0 — a major improvement on the original version (See “DrugBank Data” below) — in the 2008 database issue of Nucleic Acids Research. But GenomeQuest’s subscribers have enjoyed a preview of the 2.0 release for the past three months. This month, GenomeQuest launches DrugBank Pro, an enhanced version of 2.0 that includes investigational drug data not publicly available.

Michael McManus, GenomeQuest vice president and general manager, says  “DrugBank Pro is an important part of our offering. It wasn’t just us that noticed this was a cool database, our customers also noticed. It adds a lot of chemistry richness.” Wishart’s team was receptive when first approached by GenomeQuest. A series of meetings in Canada and Boston eventually clinched the agreement.

“We saw the [original 2006] DrugBank paper and instantly recognized it was very important for our users,” says GenomeQuest’s senior director, content development, Kamalakar Gulukota. “DrugBank may be a small database, but nevertheless it’s separating the grain from the chaff… Ultimately, it’s what all the pharma and biotech companies are looking for.”

The deal benefits both parties. In return for funding the Wishart group’s efforts (support comes also from Genome Canada), the Westborough, Mass., company gains a six-month preview of the DrugBank data such that newly approved drugs will appear in GenomeQuest before they appear on the web. Given the current pace of new drug approvals, that discrepancy won’t affect too many compounds or inconvenience users of 2.0.

More significantly, DrugBank Pro offers exclusive access to information on investigational drugs in clinical trials, which again would only appear in 2.0 once they are approved. While DrugBank 2.0 carries information on some 3,000 experimental drugs, these are ligands to proteins in the Protein Database that are not in clinical trials, and may never get there because of toxicity or other issues. 

Gulukota stresses that the collaboration helps academic and public users by helping to ensure DrugBank’s survival. “DrugBank was developed as an academic project,” he says. “David is aware that unless there’s continuous interest — someone driving it — it will fall by the wayside. Each year, Nucleic Acid Research publishes hundreds of databases, but most databases fall into disuse.”

Wishart agrees: “I think the interaction with GenomeQuest has been mutually beneficial and it certainly was key to helping get the latest release of DrugBank in tip-top shape,” he tells Bio•IT World.

Silo Mentality
The great value of DrugBank 2.0 stems from its aggregation of data that are traditionally siloed, or as Wishart terms it, “an effort to bridge the ‘depth versus breadth’ gap between clinically oriented drug resources and chemically oriented drug database.”

“The data are available in multiple sources, but you need to connect it up,” Gulukota says. “Almost all targets and drug metabolizing enzymes have multiple entries in GenBank and other databases, including patent databases. The main problem is, when you’re looking at a target, do you get the chemistry information? Conversely, when you’re looking at a chemistry database, do you get the targets?”

Gulukota continues: “There’s been this silo effect in pharma companies, where chemistry and biology don’t necessarily speak the same language, let alone talk to each other. Often you have chemists who don’t know anything about the target, [and] they don’t necessarily look at splice variants, or other receptors the chemical might bind. Many questions that could nip non-productive research in the bud are never asked because there isn’t that mindset. Pharmaceutical executives have needed to connect chemistry and biology for a long time.”

Last December, GenomeQuest offered a “sneak peek” of DrugBank Pro to its users, and hopes it will attract new users within its client base. Users will be able to ask many key questions earlier, says Gulukota: “Are there splice variants of this gene I need to be worried about? If one of those variants is in the heart, and you’re exploring neurology indications, you might want to look for cardiovascular side effects or cardiovascular indications.”

McManus notes that GenomeQuest’s high-throughput extension will be valuable in using DrugBank Pro. He also hopes that the relationship with Wishart’s group continues. “Given how creative David is, if there are other ideas he might have, we’d be open to talking to him.”