SF tech startup Atomwise gets $45M for AI-powered drug design software
A Silicon Valley software company that’s using artificial intelligence to take the guess work out of structure-based drug design is trotting out a sizable Series A round this morning. The deal includes investors that run the gamut, from tech to biopharma to agrochemicals.
The company is called Atomwise, and it’s hauling in $45 million in a round led by Monsanto Growth Ventures, tech investors DCVC (Data Collective), and B Group Capital – the VC fund founded by Facebook co-founder Eduardo Saverin.
In biotech money, $45 million is starting to seem standard if not paltry for a B round (a sign of the very flush times, no doubt). But in software, this dollar amount still makes a splash, Atomwise’s CEO and co-founder Abraham Heifets tells me.
The company is working on artificial intelligence software that will help drug designers find molecules that are worth pursuing – a lengthy process that means time and money to the industry.
A lot of companies are working to find targets on the biology side, Heifets said, asking which proteins and pathways play vital roles in diseases.
“But once you’ve answered those questions, you need a molecule that can block that protein or inhibit that pathway specifically. All of that design is on the chemistry side, and that’s what our software helps with,” Heifets said.
Right now, ultra-high-throughput screening robots do the work to sift through potential molecules hunting for the right puzzle piece. They can get through about 100,000 compounds per day. Atomwise’s software can run 10-20 million per day.
“We’ve seen a shift in the world from scarcity to abundance,” Heifets said. “We now have synthesized on-demand libraries, and you can order something like 600 million molecules on the web. But you can’t test 600 million molecules physically. To grapple with the incredible success of the pharma manufacturing community, you need computational approaches.”
Although founded back in 2012, Atomwise has seen much of its growth in the past two years. The company struck partnerships with four large pharmaceutical companies, including Merck, multiple biotechs, and over 40 major research universities.
Heifets sees the company’s tech as a solution to the pharmaceutical industry’s ROI problem.
“The industry can’t keep going with negative ROIs much longer,” Heifets said. “We need new techniques. And we think – and have the data to back it up – that AI is a solution to those problems.”
That’s a statement that might raise some eyebrows in the biopharma community. There’s a bit of a debate over whether companies like Atomwise are overstating the power of AI to save drug designers time.
“The whole compound screening step is just another early thing in preclinical space; I’ve never seen a successful project in which it was a rate-limiting step. But ‘shave a few weeks off something at the very beginning’ isn’t as compelling an offer, is it?” Lowe wrote in a blog a few months ago.
Heifets, however, is convinced. After sending Lowe’s statements to Atomwise, Heifets wrote me this in an email:
“It’s commonly understood that the Lead Optimization step is the most expensive ($414 million versus a Phase III clinical trial at $314 million, in 2010 dollars).””
Although pharma is the “bigger economic opportunity,” Heifets said the company sees customers in both biopharma and agrochemical industries.
“The computer doesn’t know or care if the carbon atom is inside a human cell or a wheat cell,” Heifets said. “The question looks the same in biochemistry. Plus, adding agrochemical compounds makes the technology more robust, so we find it valuable.”
The software was valuable enough to Monsanto, who led the Series A round.
“We chose to invest based on the impressive results we saw from Atomwise in our own hands,” Kiersten Stead, partner at Monsanto Growth Ventures, said in a statement. “Atomwise was able to find promising compounds against crop protection targets that are important areas of focus for agrochemical R&D.”
Atomwise, which employs 17 today, will use the new cash to scale, most likely doubling its staff by the end of the year and pushing the limits on the software’s capabilities.