Schrödinger, the drug discovery computing platform backed by Bill Gates, hedge fund manager David E. Shaw and Google Ventures, has closed its latest financing round at $110m, as investors back its new plan to develop its own drugs.
New investors including Invus, Pavillion Capital, a subsidiary of Temasek, and Michael Antonov, one of the co-founders of virtual reality company Oculus, have joined the fundraising round, expanding it by almost 30 per cent since the round was announced in January.
The company has until now focused on providing its technology to pharmaceutical customers. The platform was used in the discovery of two cancer drugs — ivosidenib from Agios and enasidenib, now marketed by Agios and Celgene — that have been approved by the US Food and Drug Administration.
“Our platform has been validated again and again across hundreds of targets in real-world drug discovery projects,” said Ramy Farid, Schrödinger chief executive.
The fundraising will help Schrödinger pursue its first wholly owned potential drug candidates in oncology. The company hopes to have its first potential drug candidate ready for trials by next year.
Karen Akinsanya, senior vice-president and chief biomedical scientist, joined Schrödinger last year from Merck to lead their new drug discovery programmes. She hopes the technology will significantly accelerate the process which can take several years and cost billions of dollars.
Ms Akinsanya said the emergence of large genetic data sets had opened up a “whole closet” of potential targets for the industry and that Schrödinger’s technology had created a “deep list’ of “very potent molecules” that could become drugs for the targets.
Schrödinger is one of many companies exploring the possibility that vast computing power could speed up the cumbersome and expensive early stages of drug development.
Relay Therapeutics, a Boston-based start-up, raised $400m from investors including the SoftBank Vision Fund last year, while in Silicon Valley, Insitro raised over $100m from investors including Andreessen Horowitz, and signed a deal with Gilead Sciences earlier this year.
But there have also been failures: IBM recently said it would stop selling its Watson for drug discovery artificial intelligence platform, although it will continue to service existing customers.
Schrödinger was founded in 1990, long before the latest generation of start-ups. Its approach simulates the physics of how compounds work, rather than simply using deep learning to crunch existing data looking for patterns.
Mr Farid said it is “mind-bogglingly complicated” to teach a computer the principles of physics — from the exact coordinates between every atom to how molecules behave in water.
“It was a really, really hard problem. Much harder than anyone realised. It took thousands of person years of investments from Gates and Shaw, and thousands of years of computer time,” he said. “That’s why the very simple idea of pattern recognition doesn’t work reliably.”