Immunai, Inc.

AstraZeneca already penned a AI pact with Tempus back in 2021.\n Tempus AI has signed a three-way deal with AstraZeneca and Pathos AI to construct a data-based oncology model that could ultimately be used to develop new cancer drugs.Tempus will use its “vast repository” of anonymized clinical data from cancer patients as the basis for the multimodal foundation model—a term for an AI model that can process and generate outputs from multiple data types. In return, Tempus is set to receive $200 million in data licensing and model development fees.AstraZeneca, Tempus and AI-focused biotech Pathos will all be able to use the completed model for their own R&D work. The idea is that the model can be used to “gather biological and clinical insights, discover novel drug targets, and develop therapeutics for the broader oncology community,” according to an April 23 release.“Generative AI and the emergence of large multimodal models is the final catalyst needed to usher in precision medicine in oncology at scale,” Tempus CEO Eric Lefkofsky said in the release.“Tempus has spent the last decade investing billions of dollars into collecting the necessary data needed for a foundation model of this kind to take shape,” Lefkofsky added. “We look forward to working with AstraZeneca and Pathos to apply AI-enabled solutions to advance therapies in an effort to help patients live longer and healthier lives.”AstraZeneca already penned a pact with Tempus back in 2021 to use the Chicago-headquartered company’s multimodal data to “advance novel therapeutic programs in oncology on a global scale.” Today’s announcement builds on this existing partnership, Tempus said in a statement. “Cancer drug discovery and clinical development are being transformed by the ability to analyze vast amounts of rich data using artificial intelligence,” Jorge Reis-Filho, Chief AI and Data Scientist, Oncology R&D, AstraZeneca, said in the release.“We are excited to collaborate with Tempus and Pathos to advance our data and AI-driven R&D strategy through the development of a multimodal oncology foundation model that we believe will accelerate and increase the probability of clinical success across our diverse pipeline,” Reis-Filho added.Pathos will also be involved in the new collaboration, and the biotech’s board member Mohamad Makhzoumi said the increasing role of AI in drug development means “the opportunity for companies like Pathos to build foundation models that seemed almost unimaginable a few years ago is now taking shape.”Like most Big Pharmas, AstraZeneca has plenty of AI-focused drug discovery collaborations in the works, including with Verge Genomics, Absci and Immunai. Earlier this month, Puja Sapra, Ph.D., AstraZeneca’s senior vice president for biologics engineering and oncology target discovery, told Fierce that the company has been busy building machine learning into its discovery portfolio.“Using deep screening and machine learning and AI models out there … we think will really help us accelerate drug discovery and development, but also help us go after targets that have not been druggable before,” Sapra said in an interview.

Can AI make an avatar of Mikael Dolsten? The former chief scientific officer of Pfizer will try to find out now that he’s starting a new chapter at Formation Bio after nearly 16 years at the New York pharma giant. In one of his first interviews since leaving Pfizer, Dolsten exclusively told Endpoints News that he is leading a drug-picking committee at Formation Bio, an AI-focused biopharma. In addition to chairing that group, he will serve as a strategic advisor to Formation Bio, a deeply ambitious company that has raised over $600 million, including a $372 million Series D last June . Formation Bio CEO Benjamine Liu said Dolsten brings 30-plus years of experience knowing what a great drug program looks like. He’ll help Formation Bio — which is pursuing a Roivant-esque model of picking up assets — select which drugs to in-license. But Liu said he also hopes Dolsten’s input can help teach AI models to basically imitate how Dolsten and other veteran drug hunters make these calls. “The thing we’re trying to do with Mikael is: How do we recapitulate what an AI Mikael looks like?” Liu said in an interview, saying he hopes experts like Dolsten can train the AI system how to make decisions and evaluate opportunities. Dolsten, 66, has wasted little time kicking off his post-Pfizer tenure. His last day at the drugmaker was Feb. 28, and he started working with Formation Bio on March 1. His role at Formation Bio will be announced later on Wednesday. He’s also building “Mikael’s ecosystem,” a mix of about 10 life sciences companies that he’s working with. Dolsten described his Pfizer tenure as “a fabulous run exceeding my wildest aspirations,” adding that he started to plan his departure about two years ago. Dolsten’s exit was first announced in July, and he helped choose Chris Boshoff as his successor. Leaving Pfizer was his own decision, he said, and wasn’t influenced by Starboard Value’s activist campaign. All told, Dolsten brought more than 35 products through approval, including the Covid-19 vaccine, Comirnaty. “I just felt that it would be hard to beat a lot of those amazing things we accomplished,” Dolsten said. “I felt intrigued by working in a new environment.” Dolsten’s other roles include board seats at the instrument maker Agilent Technologies, the New York AI biotech Immunai , and the Danish macrocycle biotech Orbis Medicines . He said he is also helping launch a new biotech startup with a serial entrepreneur from Yale University. That said, he’s spending relatively more time on his Formation Bio role, estimating he spends about a day per week at the company. Also on Formation’s drug-picking committee is its newly-hired chief business officer David Steinberg, a co-founder and former executive at PureTech Health. Steinberg joined in October. Dolsten described his conversations with Liu over the past few months as a key factor in his decision to join the company. He described Liu as a “high-energy AI magician with a vision.” One of Formation’s newest projects is using AI, particularly the latest generation of deep research tools , to create an atlas of all possible drug targets. This includes scouring press releases, scientific conference abstracts, company websites and more to aggregate up-to-date reports for each target. The goal, Liu said, is an “equivalent of RA Capital’s atlas , but instead of 50 PhDs, we have AI systems go through target by target.” With about 150 employees right now, Formation Bio is betting that AI-based workflows can be more efficient than a 100,000-person pharma company. Dolsten said he’s impressed by what today’s AI systems can produce, compared to the conventional way that large drugmakers like Pfizer and AstraZeneca operated while he worked there. “So much of that work required advisory committees, multiple business development meetings, and now can happen in the instant of a moment,” Dolsten said, estimating that about 80% of that work can happen in 30 minutes using these AI systems.

Through its re-upped pact with AstraZeneca, Immunai is slated to receive $18 million for the first phase of the new project. AstraZeneca is extending its work with Immunai through a multiyear collaboration that aims to implement its artificial-intelligence-powered immune cell atlas in the design of the drugmaker’s clinical trials of cancer immunotherapies.The two companies previously partnered on research into targets within inflammatory bowel disease based on cell-by-cell analyses of the immune system.Immunai had also lent its multi-omics tech to the development of AstraZeneca’s CTLA-4 and PD-L1 bispecific antibody volrustomig, currently being tested in phase 3 studies among cancers of the lung, cervix, head and neck. Founded by former researchers and computer engineers from MIT, Harvard and Stanford University, New York-based Immunai has been building a digital map of the immune system by charting each cell type’s function and their interactions with the body and various diseases.In late 2021, the startup collected a mammoth $215 million series B round after raising $60 million through a series A earlier that same year. Those funds have helped the company stretch its reach from immuno-oncology toward broader research into auto-immune, cardiovascular and neurological conditions both through its AMICA cell atlas and its AI model dubbed the Immunodynamics Engine, or IDE. Through its re-upped pact with AstraZeneca, Immunai is slated to receive $18 million for the first phase of the new project.The collaboration as a whole will focus on identifying which patients may be most likely to respond to different treatments based on profiles of their immune systems and the drug’s mechanism of action—while also spanning clinical decision-making, dose selection and biomarker identification, according to Immunai. The Big Pharma will then have an option to further expand the teams’ remit.“This collaboration is a natural progression of our successful work in advancing drug development in the fields of oncology and immunology,” Immunai CEO Noam Solomon, Ph.D., said in a statement. “Bringing a drug to market is incredibly challenging, time consuming and expensive. Through this collaboration with AstraZeneca, we’re excited to leverage our AI-based engine, the IDE, to help make this process more efficient in bringing potential new therapies to patients.”