Vancouver-based Absci presented its newest machine learning models this week, displaying a computer-generated model that can predict — with a high degree of confidence — a drug candidate’s developability.
“Our (artificial intelligence) can predict the best drug candidate with the functionality, the developability and manufacturability that we want,” Sean McClain, founder and chief executive at Absci, told The Columbian.
McClain said computers can take the scientists developing the drugs significantly farther through the process of creating a drug candidate. And the formula determined by the computer has an extremely high confidence — about 85 percent. Then it’s tested in the lab.
“That means that drugs are going to get to patients a lot sooner than previously,” McClain said. The streamlined process, he added, will also increase the probability of success.
The process for creating a drug can be time consuming. There are more options to create a drug candidate than there are atoms in the universe, McClain said.
“So the (artificial intelligence) is allowing us to fully — on a computer — be able to search through all of that and figure out for a particular disease, what is the best drug to design,” he said.
“It’s a huge, huge, huge breakthrough,” he added. The results are possible because of the technology that creates the data that can in turn be used by artificial intelligence.
Absci is working with NVIDIA, an interactive graphics company, to advance Absci’s drug pipeline with NVIDIA software and hardware and both Absci and NVIDIA technical expertise to expand and scale the machine learning models.
The Vancouver-based company presented their findings at NVIDIA GTC, a global conference on artificial intelligence.
“The huge breakthroughs that we’ve had are continuing to drive our business development and partnership pipeline,” McClain said. “It’s going to result in more and more active programs, which translate into long term value creation.”
The company also announced its 2021 financial and operating results this week, a requirement for public companies, which Absci has been since last summer.
The announcement detailed that the company created five biologic drugs using its technology last year and three so far in the first three months of 2022. Absci also partnered with Merck, a large pharmaceutical company, to create up to three drugs.
McClain said Absci’s key metrics are the number of active programs, or created biologic drugs, it has. The team met the goals for five active programs last year and expects to meet eight this year.
“We have a very strong pipeline,” McClain said.
The financial picture wasn’t totally rosy for the company, as it reported a $101 million net loss last year, compared with a loss of $14.4 million in 2020. The report also noted Absci expects a net decrease in cash, cash equivalents and restricted cash of $120 million in 2022.
“We expect to continue to lose capital because we’re making really important investments in (artificial intelligence),” McClain said. “The huge accomplishments that we’ve had in the space wouldn’t happen unless we’re making investments in those.
“We’ve made it very clear to investors that we’re going to continue to make these investments,” he added. “We’re going to continue to burn capital, but we’re seeing the impact of that and it’s a very big net positive for the long term.”
McClain called the story of the company “long term.” Absci’s investors are supportive of that, he said. The cash burn and revenue are both in line with what analysts had projected, he added.
Absci trades on the Nasdaq under the ABSI ticker symbol. Share prices were up 5 cents Wednesday to $7.53.
This article was corrected to list the number of drugs Absci will be making for Merck, as of it’s January partnership.
