Some new investors pitched in to Recursion Pharmaceuticals Inc.'s $121 million series C financing, adding to the industry's vision of the importance of AI and machine learning.
The round was led by U.K.-based Ballie Gifford's Scottish Mortgage Investment Trust plc, but new institutional investors since Recursion's 2017 $60 million series B include Intermountain Ventures, Regents of the University of Minnesota and the Texas Tech University System.
"They have invested at the intersection of biology and technology, and they have us as an anchor in both," Recursion's CEO, Chris Gibson, told BioWorld.
Gibson said he sees the place of AI and machine learning-based companies such as Recursion as changing rapidly. If a company, such as Recursion, which leverages AI to discover new therapies, wants to succeed, they have to rethink not only their staffing but how the staff interacts within the company, Gibson said.
"I recently have had discussions with many C-level executives from top 10 pharmaceutical companies. I noticed how our teams are massively cross-functional. This is different from other companies. Very often they've been in silos," Gibson said. "Here, a biologist calls up a data scientist to ask questions. All the stakeholders help decide how to ask questions, all areas are involved."
Recursion recently appointed Sharath Hegde, as its chief scientific officer. Hegde has more than 28 years of leadership experience in drug discovery and clinical development and has seen multiple drugs from discovery through FDA approval. He joins the company after more than 15 years at Theravance Biopharma Inc.
The task for AI-machine learning companies, Gibson said, is to think into the next five or 10 years instead of succumbing to the near-term thinking that dominates so much of the biotech industry and Wall Street.
"My strong sense is that, independent of recruiting outcomes, you have to look long term. We were founded based on a hypothesis that the application of technology to biology would require insights from a large swath of people," Gibson said. "The founders of Recursion are a computer scientist, a genetics guy and a business guy. We very much look like that as a team. The question is, can more established companies recreate that? Or partner with companies like us? It is excruciatingly difficult to recreate that because public company boardrooms feel pressure to jump on near-term decisions."
The new financing will support the build-out of Salt Lake City-based Recursion's machine learning-enabled drug discovery platform. It also allows the company to advance its preclinical and clinical assets, including clinical-stage programs for cerebral cavernous malformation and neurofibromatosis type 2. The company also plans to continue to forge partnerships with pharmaceutical companies in a variety of therapeutic areas, including immuno-oncology, oncology, aging and inflammation. Gibson said he expects to announce new partnerships in those areas in the coming year.
All of Recursion's prior institutional investors also participated in the round, including Lux Capital, Data Collective, Mubadala Ventures, Two Sigma Ventures, Obvious Ventures, Felicis Ventures, Epic Ventures, Menlo Ventures, AME Cloud Ventures and CRV.
Recursion was founded in 2013. While its work typically starts with a fairly traditional approach to drug discovery, phenotypic screening, it then layers on some of the most transformative elements of modern computing, including machine learning and artificial intelligence, in an effort to yield deeper insights than phenotypic screening alone. (See BioWorld, Dec. 29, 2015.)
The company models thousands of diseases in human cells, capturing about 2 million new fluorescent micrograph images every week across about 90,000 different mini-experiments, a process by which it is building one of the world's largest biological image datasets. The efforts are all aligned toward creating what Gibson calls a "highly relatable and comprehensively annotated dataset," which can be used to train computational neural networks and to identify changes in thousands of cellular and subcellular features associated with various diseases.
When the process uncovers a drug that appears to make a diseased-looking cell look healthy again, it is short-listed. The reins are then handed over to a team of biologists who use all the same traditional approaches that typical biopharma ventures employ to understand whether or not a candidate has value. (See BioWorld, Oct. 23, 2017.)
In May, Recursion announced that it will open-source a glimpse of the biological dataset the company has been building for more than five years. At more than 2 petabytes, and across more than 10 million different biological contexts, Recursion labels its data as the world's largest image-based dataset designed specifically for the development of machine learning algorithms in experimental biology and drug discovery.