AI startup Entos looks to one-up the rest of the field by folding quantum mechanics into its discovery platform
Imagine predicting molecular properties 1,000 times faster with 100 times less training data. That’s the future that co-founders Tom Miller and Fred Manby envision at Entos. On Wednesday, the team unveiled a $53 million Series A round and a new 16,000 square-foot San Diego headquarters to get started.
Entos came together last April with tech derived from Miller and Manby’s labs at Caltech and the University of Bristol, respectively. The two academics had been operating in the same circles for a while before they officially met at a conference in East Lansing, MI.
“Over the course of that year, it’s just been an explosion of progress. It’s been incredibly exciting,” Miller, now CEO, told Endpoints News, adding that the team has grown from two to 30.
Their platform, dubbed OrbNet, uses quantum mechanics to encode and map chemical space. In addition to assisting with hit identification, Miller says the technology could be used for “those very costly, and therefore very valuable later stages” associated with lead optimization and getting a drug all the way to the clinic. And they’ve convinced Frances Arnold — who won the Nobel Prize a couple years back for her work on the evolution of enzymes — to sign on as a scientific advisor.
“The challenge of any machine learning application is to go from a molecular structure to a particular property,” he said. “So that is a mapping problem and OrbNet uses a representation that is fundamentally connected to quantum mechanics that leads to far greater efficiency in terms of the required data to accurately perform that mapping, and far greater transferability through chemical space.”
The Series A funds will allow “rapid progress” in the company’s hiring activity, Miller said, and has already helped fund its new San Diego laboratory. While the company has a “growing portfolio” of partnerships in the works, the focus is on internal development programs, the CEO added.
The machine learning space is packed with players looking to revolutionize the way new drugs are found and developed. Last month, Insilico’s Alex Zhavoronkov pulled in a $225 million Series C round. Exscientia raised a fresh half-billion dollars for its AI platform and pipeline earlier this year. And in March, Daphne Koller hooked $400 million for her machine learning startup, insitro.
“Drug development is a highly complex endeavor that is plagued by both cost and time inefficiency, and we believe that incorporating Entos’ impressive technology into the process will result in faster timelines and vastly improved therapeutics,” said Aaron Weiner, managing director and head of healthcare at Coatue, which co-led Entos’ Series A round with Catalio Capital Management.
OrbiMed, Sequoia Capital, Nexus Ventures and Freeflow also contributed to the round.