Compugen wants to dip its toes into the pharmacogenomics pond.
Two years ago, the Israeli bioinformatics company began laying the groundwork for a vast and comprehensive data repository it hopes drug makers will find irresistible. This summer, Compugen plans to commercialize it as a service that ties together an array of clinical data from millions of patients that biopharmas can use to improve drug-development protocols.
“The data [Compugen will offer] is very scarce,” according to Michal Preminger, vice president of the new research direction. In the United States, she said, “most patients’ lives are not computerized at all.” So the firm sought salvation in European and Israeli health-care providers.
Actually, the data are just one part of a platform that Compugen hopes will generate revenue and, eventually, provide a base on which to develop its own drugs.
The service, which is part of a new group within Compugen called Predictive Drug Response, was envisioned to become a kind of clinical sounding board for biopharmas. The data comprise 15 years worth of records and independent physician opinions concerning diagnoses, lab and limited genetic tests, various clinical images, prescription histories, and hospitalizations culled from some five million patients.
In principle, biopharmas may use these data together with the SNPs and haplotypes in Compugen’s LEADS platform to decide which individuals to enroll into clinical trials, to invigorate mediocre drugs in phase IV studies, or to revive drug candidates with less than favorable adverse-event profiles. And though a handful of existing companies now occupy a sliver of this space — Genaissance applies its haplotype database to help pharmas chase personalized medicine, and bioinformatics firm AnVil partners with health-care companies to mine and analyze data for pharmas — Compugen’s stance is unique thanks to the breadth of its database and the ancillary information in its LEADS platform.
“Compugen … would have naturally gone into the direction of pharmacogenomics, but given the fact that our core competence is in taking computational technologies and merging them with the [other] disciplines, I thought we could come up with a stronger advantage,” said Preminger, who spearheaded the project. “I know there’s a good opportunity for service there.”
As it stands now, the platform, which includes data from common ailments like diabetes and heart disease, has four components: The first two are patient data and data analysis. Because they are not culled from trials, these data are unencumbered by the narrowly focused clinical information currently used by drug makers, said Preminger. “It’s real people in real life,” she said. “There’s no exclusion/inclusion criteria.”
“Our goal is to combine all of this data … with certain molecular analyses that will complement the picture and come up with reliable predictions about things like drug response and disease progression,” Preminger said. “The heart of this is to predict a patient’s response or disease progress based on history.”
To arrive at their data, Compugen analyzes the patient medical records and creates a homogeneous sub-population characterized by a particular response — for example, drug resistance, non-responders, and adverse reactions. Drug companies can then use the product of these analyses to better select patients for certain studies or make early predictions. Saving money on costly late-stage clinical trials is a drug maker’s holy grail, and knowing which patients to enroll is one significant way to save tens of million of dollars per drug candidate.
The remaining two components are computational molecular analysis and experimental analysis. In the former, Compugen will use technologies in its computational biology platforms, especially from the SNPs and haplotypes found in its LEADS platform, “to design and analyze molecular experiments” to extract additional predictive information. In the latter, which is still under development, Compugen would search for “novel methodologies,” in particular protein analysis.
In the long run, though, Preminger, who works in Compugen’s Jamesburg, NJ, facility, said that the first two points “are our strongest. If we didn’t have that then we would probably remain unwilling to enter this field at all.”
“When we decided to get into this field, we felt that there were two good reason for that,” she told SNPtech Reporter over lunch this week. “With our combined business model … we thought it was a good idea to provide services that are more downstream. Our traditional markets are in the discovery side, and we felt it was very important for us to be able to provide services which are closer to the end product, and which allow the benefits to be presented in the shorter term than in the discovery phase.”
Preminger said that Compugen has so far conducted one project with a pharmaceutical collaborator. Additionally, she said other pharmaceutical companies have “shown interest” in the platform, especially firms with an interest in cardiovascular diseases and diabetes.
Though she said it would be difficult to determine the size of this market — the concept is still new — Preminger said Compugen initially “will be happy to offer this service because we will learn a lot from it. We can definitely use the revenue. But I think longer term we would try to use this platform for our own drug discovery or for other types of collaborations.”