Bringing big data to drug discovery
twoXAR, co-founded by MIT Sloan alumnus, uses machine learning to identify drug candidates in minutes, not years.
By Brian Eastwood |
September 28, 2016
Andrew M. Radin, MBA ’14, co-founder and chief business officer of drug discovery company twoXAR
Many biopharmaceutical firms talk about the potential of big data to change the drug discovery process, but a Palo Alto, Calif.-based startup co-founded by an MIT Sloan alumnus is doing it—and has completed scientific research studies to prove it.
Founded in 2014, twoXAR (pronounced “two czar”) analyzes large biomedical datasets to find needles in a haystack: the handful of drug candidates out of a drug library of thousands, if not millions, that would most effectively treat a specific disease.
“Leveraging our technology, we should be able to shave years off of the standard drug development process that sometimes takes 15 years and costs billions of dollars,” said Andrew M. Radin, twoXAR co-founder and chief business officer. The company recently completed a preclinical study of candidate drugs to treat rheumatoid arthritis and has since started research on a type of liver cancer known as hepatocellular carcinoma, or HCC, in collaboration with Stanford University.
“We use computation to screen large libraries of molecules in a matter of minutes against millions of data points spanning chemical, biological, and clinical databases,” said Radin, who received his MBA from MIT Sloan in 2014. “Analyzing data isn’t new, it’s a bedrock of the scientific method, but using computers to take a big data approach to identifying new medical treatments is finally a reality.”
Augment, not automate
Three factors have enabled twoXAR to build its business, Radin said: The emergence of cloud computing, advances in machine learning, and the amount and diversity of publicly available structured data from organizations such as the National Institutes of Health and research institutes in Europe and Japan.
Co-founder and chief executive officer Andrew A. Radin (no relation) developed the idea for twoXAR’s approach for analyzing large sets of biomedical information while pursuing graduate coursework in the biomedical informatics program at Stanford University.
These algorithms drive the company’s drug discovery platform, which finds unanticipated associations between particular diseases and drugs that might be able to treat them. The platform then analyzes each of the identified candidate drugs to determine how effectively they could treat that disease.
Andrew M. Radin said the platform is meant to augment, not automate, the drug research and development process. Instead, it speeds up the early, discovery stage of the process. This echoes the principles of “racing with the machines” spelled out in The Second Machine Age, co-written by Erik Brynjolfsson, one of Radin’s professors at MIT Sloan.
While machines can tease out hidden associations in data, this is not the same as human insight, Radin said. Humans are great at hypothesizing about and understanding the rationale behind the associations identified by the computer. That’s where the expertise of life scientists comes into play.
“Computer scientists and life scientists need to come together as researchers to move medical treatment hypotheses forward quicker than we, as society, have been using traditional approaches,” Radin said. “Computation enables a speed and scale of analysis previously unavailable to medical research and we have done this in our partnerships with disease researchers; we’ve used our software platform to identify numerous new ideas across therapeutic areas and validate them in standard assays in a matter of months.”
This approach to the market is part of what makes twoXAR stand out, said MIT Sloan Professor Scott Stern. Radin took three courses with Stern, including a doctoral-level course on economic innovation and entrepreneurship, and both co-founders have met with Stern for strategy sessions.
“They have been deliberate in how to establish value for customers. They’re not doing science for science’s sake but making decision-making tools for their customers,” Stern said. “What I liked from the very beginning was that they had a clear, underlying hypothesis about how they would create value. They’re taking an abstract idea and turning it into a concrete implementation.”
In November 2015, twoXAR raised $3.4 million in seed round funding from the new $200 million Andreessen Horowitz “bio fund.” Their other investors include CLI Ventures and Stanford’s startup incubator, StartX.
As twoXAR has grown, Radin has learned that deal-making in a startup is often about selling innovation just as much as it is selling a specific product or service. Here, he said, he draws on a key lesson from MIT Sloan: The importance of communication. While twoXAR focuses on math and computer science, potential clients such as pharmaceutical companies focus on chemistry and biology. “When communicating with our audience in pharmaceutical companies, we focus on results from preclinical studies based on predictions from our technology and how we can leverage our software platform to help them find the next blockbuster drug,” Radin said.
Leveraging the MIT community
The relationships Radin built during and after his time at MIT Sloan have been valuable to twoXAR. In particular, the company has made industry connections through the MIT Center for Biomedical Innovation, appeared at the MIT Industrial Liaison Program’s Digital Health Conference last fall, and remains a member of that program’s MIT Startup Exchange.
While this type of access is not unique among business schools, Radin said MIT Sloan offers an environment that’s more cooperative than competitive.
Stern agreed, saying the MIT community likes to see companies like twoXAR succeed. “There’s not enough time to build all the great companies that we can, so there’s a real commitment by faculty, students, and the campus to nurture new ventures,” he said.