Immunai, a startup developing a proprietary corpus of immune-centric human tissue analyses, today emerged from stealth with $ 20 million. It will use the capital to further the development of its tech and business functions while hiring new scientists, engineers, and machine learning experts, a spokesperson told VentureBeat.
Emerging treatments like gene cell therapies and cancer immunotherapies promise to revolutionize the field of medicine. But the immune system’s complexity — trillions of cells partitioned into hundreds of types and states that interplay with various systems and proteins — threatens to stymie research. In 1999, a patient in a trial died after an immune system attack likely resulting from preexisting antibodies against a virus used as part of gene therapy — a death that experts believe led to years lost in gene therapy development. Immunai aims to prevent such mistakes with immune profile-generating AI.
Immunai was founded in December 2018 by CEO Noam Solomon, an ex-Harvard and -MIT postdoctoral researcher, and CTO Luis Voloch, a fellow MIT graduate and former machine learning engineer at Palantir. The two teamed up with members of the Parker Institute, which works with researchers to accelerate the development of immune therapies — Danny Wells, a founding data scientist, and Ansuman Satpathy, a cancer immunology professor — to pursue a platform that sheds light on cell populations post- and pre-treatments.
Immunai’s tech derives over a terabyte of data from a single blood sample, profiling cells at what the company claims is “unprecedented” depth and scale. Samples are compared with a database using AI that maps data to hundreds of cell types and states, creating immune profiles based on highlighting different elements and characteristics.
It’s an approach similar to that of scientists affiliated with the Human Vaccines Project, who are working to identify biomarkers (i.e., indicators of particular disease states) that predict immune responses to vaccines and cell therapies. Elsewhere, Microsoft and startup Adaptive Biotechnologies are collaborating to develop algorithms that create a “translation map” for cell receptors to antigens — pathogen molecules that trigger an immune response — and map those antigens back to diseases.
Clinical studies have traditionally focused on testing thousands or even tens of thousands of subjects and collecting a limited amount of data on each, but massive corpora and AI enable millions of data points to be collected about a single individual. It’s Solomon’s belief this will better reveal the mechanisms underlying diseases. “When looking at only a specific disease or patient cohort, one gets a limited and siloed view of the immune system,” he said. “By using machine learning and applying it to our proprietary diverse database of single-sequencing data paired with rich clinical data, our platform identifies common patterns that are not visible when looking at the narrower disease-specific view.”
The profiles could support the discovery of novel biomarkers by spotting changes in cell type and state-specific expression. In a peer-reviewed study about the inhibition of a protein known as programmed cell death protein 1 (PD-1 and CD279), which sits on the surface of cells and has a role in regulating the immune system’s response, the Immunai team uncovered information about the origin of tumor-fighting T cells.
“Our mission is to map the immune system with neural networks and transfer learning techniques informed by deep immunology knowledge,” said Voloch. “We developed the tools and knowhow to help every immuno-oncology and cell therapy researcher excel at their job and increase the speed in which drugs are developed and brought to market, while making sure impacts are understood.”
Immunai is headquartered in New York City, with offices in San Francisco and Tel Aviv. In addition to Satpathy and Wells, it counts Dan Littman, a New York University professor of molecular immunology, as its third founding scientist.
Viola Group and TLV Partners led the seed funding round.