Start Your 'Innovation Engine': Software Predicts Drug Winners
Betting on the wrong drug can easily cost millions of dollars in wasted funding from private investors or American taxpayers. Now a startup's "Innovation Engine" software could slash the risks by predicting the biotech inventions that are most likely to succeed as both medical therapies and marketable products.
The program could help government labs and pharmaceutical companies avoid wasting time and money in chasing down dead-end drugs or therapies . It could also give a much-needed injection of optimism during a time when pharmaceutical companies have begun panicking about their failure to find new drugs a concern for patients as well as profit margins.
The Innovation Engine was born from the frustrations of two Ph.D. students at Tufts University near Boston. They tried to commercialize several biomedical technologies, but soon realized that universities and biotech companies faced a bewildering array of technologies to gamble upon.
"It's an inefficient process," said Dave Greenwald, co-founder and CEO of Relay Technology Management. "People in this industry are very passionate about doing the right thing, and they just need a tool to make the right decisions."
Pharmaceutical giants such as Sanofi-Aventis and Pfizer have cut back sharply on their own spending to find new drug candidates, said Kenneth Kaitin, director of the Tufts Center for the Study of Drug Development and an advisor for Relay's co-founders. He added that the industry is looking to outside sources for innovation.
"I think it is panic time for the industry," Kaitin said in a phone interview with InnovationNewsDaily. "They're gambling that they'll be able to find new assets new drug candidates and replace some of their research activities with collaborations involving small companies and academic institutions."
Figuring out innovation
Gauging the measures of successful innovation has proven tricky, even for experts. But the Innovation Engine works by collecting clinical trial and market data from both public and private sources, including 500 academic research centers and 2,500 biotechnology companies.
Relay's algorithms then run analyses to weigh how certain factors such as intellectual property, commercial potential and scientific standing contribute to innovation.
"We're identifying what's up and coming, what's next, and presenting it in a data-driven fashion that takes into account not just the science, but also what market is doing," Greenwald explained. "It's kind of predicting the behavior of the pharmaceutical industry."
Greenwald, who obtained his Ph.D. in genetics at Tufts University, hit upon the possible solution with Brigham Hyde, a Ph.D. in pharmacology. They also enlisted the help of Rachel Lomasky, a Ph.D. in computer science who specialized in so-called machine learning.
Paying users can access the Innovation Engine to search for the most promising new drug candidates for a number of diseases or conditions. The Web-crawling software responds by providing a ranking of the best candidates.
Standing tests of time
Relay also verified its method by pitting its software's predictive power against known history. The Innovation Engine successfully spotted six of 10 big licensing opportunities seized by biotech companies during one test analysis that took place over eight months.
Similarly, it predicted many of the blockbuster drugs from the past 50 years, such as Pfizer's cholesterol-lowering drug Lipitor.
The Innovation Engine also surprised the Relay co-founders by dispelling certain assumptions regarding what might lead to a successful medical innovation. Getting published in prestigious journals such as Science or Nature or being developed at top universities such as MIT or Caltech proved no guarantee of such success.
As it turns out, the rules of success seem to change for every therapeutic category, Greenwald said. That means different factors matter for diabetes treatments as opposed to cancer therapies .
The ability to spot trends in the biotech industry can help pharmaceutical companies figure out what to license, but it can also help universities decide which companies they should approach with new drug candidates.
The best and brightest
Newer versions of the Innovation Engine can even single out which research labs or individual researchers have proven the most productive in creating innovations , or what their areas of strength are concerning specific categories of innovation.
"It could be used for granting agencies such as the [National Institutes of Health]," Greenwald said. "They've made hundreds if not thousands of decisions about which labs are funded. We can look at the funding decisions and see how successful these efforts were in generating therapies."
Relay has already had paying customers in the form of biotech companies, and has talked to private investors about the first round of venture capital funding. All of that goes toward the company's goal of not only streamlining existing medical innovation efficiency, but also finding great new innovators or innovation sources hidden under the radar.
"It's meant to be something provocative to some extent," Greenwald said. "What we're hoping is to find the really good researchers which [the industry] hasn't heard about. There's really good science going on in many places, not just in D.C. or San Francisco."
Relay Technology Management represents one of 15 student teams or startups showcased in the Open Minds competition hosted by the National Collegiate Inventors and Innovators Alliance in partnership with the Lemelson Center for the Study of Invention and Innovation. The public event took place at the Smithsonian's National Museum of American History in Washington, D.C., on March 26, 2011.