Numerate uses Google-like code to develop drugs


Using algorithms, cloud computing and other high-powered tools, a handful of Bay Area companies is sorting through millions of compounds faster than ever before to design, create or redirect compounds, then decide whether to test the drugs in humans.

Rational drug development has long been the Holy Grail of biopharmaceutical companies — since each step in the process raises the stakes by using more time, patients and money — but with minimal success. Yet with development costs rising and patents waning, Big Pharma’s interest is increasing. That is feeding a new generation of companies with more powerful computer technology at its fingertips.

The ultimate goal, however, remains the same.

“At the end of the day, we deliver white powder in vials,” said Guido Lanza, president and CEO of 17-employee Numerate Inc.

Created three years ago from technology out of now-defunct Pharmix Corp., San Bruno-based Numerate uses a recently patented algorithm that works much the same way that Google Inc. uses an algorithm to rank its search engine results. Instead of listing popular Lindsay Lohan web references, however, Numerate uses its calculations to find, structure test and ultimately make compounds for companies.

Meanwhile, startup SeaChange Pharmaceuticals Inc. is using a new method — called SEA, for “similarity ensemble approach” — to identify drugs that can be used against diseases outside the realm of their original targets.

The company was founded by Michael Keiser, Brian Shoichet and John Irwin out of the University of California, San Francisco, and is based in the Garage incubator at UCSF’s QB3 institute at Mission Bay.

It’s more than a computational advance — it’s a new way of thinking

Drug companies have shifted their drug-development philosophy from a one-target, one-drug “magic bullet” approach, he said, to a “magic shotgun” where one drug can target many diseases.

Still, there are limitations.

After Numerate discovered and produced hepatitis C-fighting compounds for San Francisco’s Presidio Pharmaceuticals Inc., the drugs were “so hard to make,” Presidio CEO Omar Haffar said.

In the end, Presidio didn’t pursue the Numerate-developed compounds because they were not vastly different from competing ones that are three or four years ahead in the process and Presidio needed to focus its cash on its main programs.

“The technology is, frankly, very exciting, very interesting,” Haffar said. “They’re really very smart people.”

Numerate, meanwhile, says it is taking the process a step further. It expects government funding soon that will predict where a drug will go in the body and how it will interact with other molecules. That could play an important role in not only predicting a drug’s success in knocking out cancer cells, for example, but also whether the drug will damage nearby healthy cells.

“The next chunk of computer power will allow us to spend more time on each compound,” Lanza said.