Immunai, an AI biotech company specializing in mapping the human immune system, and the Parker Institute for Cancer Immunotherapy (PICI), a San Francisco-based nonprofit organization that unites leading immuno-oncology experts from cancer centers across the United States, have announced a new partnership to build “one of the largest” single-cell datasets derived from a single cohort of patients treated with standard-of-care immunotherapy.
Immunai will perform single-cell RNA sequencing (scRNA-seq) and multi-omic profiling of patient blood samples collected from PICI’s Resistance Drivers for Immuno-Oncology Patients Interrogated by Harmonized Molecular Datasets (RADIOHEAD) cohort, a prospective longitudinal study of 1,070 patients receiving immune checkpoint inhibitor treatment regimens in a community setting. The cohort is pan-cancer, including samples from patients with non-small cell lung cancer, small cell lung cancer, gastric cancer, hepatocellular carcinoma, melanoma, and others.
According to Noam Solomon, PhD, CEO of Immunai, pursuing a database that contains more than 1,000 patients at different time points with various indications is “definitely one of the largest” cohorts that have been profiled with single cell resolution, given that limitations in resources often prevent investment in collecting large patient course data and the intensive technology pipelines needed for deep genomic or proteomic profiling.
At the core of Immunai’s platform is its Annotated Multi-omic Immune Cell Atlas (AMICA), an immunology database comprised of high-resolution single-cell data from public and proprietary clinical cohorts and experiments across tens of millions of cells. Data generated from RADIOHEAD will be integrated into AMICA to support comparative analysis across diverse cancer types.
Solomon said the resulting dataset will be made publicly available to the biotech community for broad use across academia and industry.
“One of the biggest problems in biology is that data is not shared for various commercial reasons,” Solomon told GEN. “The intent was to share the data from the beginning. If pharma companies want to use the data for doing discovery, I think it’s going to be great.”
“Our mission to turn all cancers into curable diseases starts with understanding why treatments work for some people and not for others. Our collaboration with Immunai gives researchers a powerful resource to decode immune responses in real-world patients—and ultimately, to bring better immunotherapies to the people who need them most,” said Tarak Mody, PhD, chief business officer at PICI.
One notable application of a broad pan-cancer dataset is the ability to identify both similarities and differences in the response and resistance to therapies across different cancers, a “key” question in systems immunology. As an example, melanoma and non-small cell lung cancer are two indications where patients are treated with an anti-PD-1 therapy, a type of immunotherapy that uses antibodies to block a checkpoint protein on T-cells which then inhibits cancer cells.
“You can imagine that a patient with a melanoma tumor and a patient with a non-small cell lung cancer tumor will not have the exact same response because they don’t have the same disease,” Solomon told GEN. “This RADIOHEAD cohort is very interesting because it enables, within cancer and solid tissues, investigation into these prominent indications to try to unlock resistance.”
Solomon emphasizes that AMICA is the company’s differentiating factor that is growing exponentially. The database currently consists of about 100 million cells and approximately 20,000 different patient samples, which is a “very good beginning” after the founding of the company in 2018. Solomon states the vision is to scale up by a 50x factor in a few years to reach a patient cohort of a million.
Additionally, Immunai’s AI algorithm, termed Immunodynamics Engine (IDE), allows the company to extract mechanistic immune insights and guide decision-making in drug development and clinical trials. IDE is composed of two AI models: AMICA ONE, a transformer-based model that pinpoints correlations of biological interest and makes predictions, such as identifying toxicity or optimal dosing, and AMICA REASON, a module that provides explainability of insights, which Solomon describes as an area of high investment for Immunai in 2025 and 2026.
“We believe it’s not enough to get good predictions. Explainability in biology, and in immunology in particular, is complicated,” said Solomon. “When you have an insight that you can explain, that’s when people are going to listen and follow the recommendations.”
Taken together, Solomon views the PICI collaboration as an opportunity to position AMICA as a valuable tool for Immunai’s partners.
“I hope that by sharing what we can find with AMICA ONE and AMICA REASON on the RADIOHEAD dataset, the community will get a glimpse of the platform and work with us even further,” said Solomon.
The post Immunai and Parker Institute to Build Massive Cancer Single-Cell Dataset appeared first on GEN - Genetic Engineering and Biotechnology News.
Immunai will perform single-cell RNA sequencing (scRNA-seq) and multi-omic profiling of patient blood samples collected from PICI’s Resistance Drivers for Immuno-Oncology Patients Interrogated by Harmonized Molecular Datasets (RADIOHEAD) cohort, a prospective longitudinal study of 1,070 patients receiving immune checkpoint inhibitor treatment regimens in a community setting. The cohort is pan-cancer, including samples from patients with non-small cell lung cancer, small cell lung cancer, gastric cancer, hepatocellular carcinoma, melanoma, and others.
According to Noam Solomon, PhD, CEO of Immunai, pursuing a database that contains more than 1,000 patients at different time points with various indications is “definitely one of the largest” cohorts that have been profiled with single cell resolution, given that limitations in resources often prevent investment in collecting large patient course data and the intensive technology pipelines needed for deep genomic or proteomic profiling.
At the core of Immunai’s platform is its Annotated Multi-omic Immune Cell Atlas (AMICA), an immunology database comprised of high-resolution single-cell data from public and proprietary clinical cohorts and experiments across tens of millions of cells. Data generated from RADIOHEAD will be integrated into AMICA to support comparative analysis across diverse cancer types.
Solomon said the resulting dataset will be made publicly available to the biotech community for broad use across academia and industry.
“One of the biggest problems in biology is that data is not shared for various commercial reasons,” Solomon told GEN. “The intent was to share the data from the beginning. If pharma companies want to use the data for doing discovery, I think it’s going to be great.”
“Our mission to turn all cancers into curable diseases starts with understanding why treatments work for some people and not for others. Our collaboration with Immunai gives researchers a powerful resource to decode immune responses in real-world patients—and ultimately, to bring better immunotherapies to the people who need them most,” said Tarak Mody, PhD, chief business officer at PICI.
Explainability is complicated
One notable application of a broad pan-cancer dataset is the ability to identify both similarities and differences in the response and resistance to therapies across different cancers, a “key” question in systems immunology. As an example, melanoma and non-small cell lung cancer are two indications where patients are treated with an anti-PD-1 therapy, a type of immunotherapy that uses antibodies to block a checkpoint protein on T-cells which then inhibits cancer cells.
“You can imagine that a patient with a melanoma tumor and a patient with a non-small cell lung cancer tumor will not have the exact same response because they don’t have the same disease,” Solomon told GEN. “This RADIOHEAD cohort is very interesting because it enables, within cancer and solid tissues, investigation into these prominent indications to try to unlock resistance.”
Solomon emphasizes that AMICA is the company’s differentiating factor that is growing exponentially. The database currently consists of about 100 million cells and approximately 20,000 different patient samples, which is a “very good beginning” after the founding of the company in 2018. Solomon states the vision is to scale up by a 50x factor in a few years to reach a patient cohort of a million.
Additionally, Immunai’s AI algorithm, termed Immunodynamics Engine (IDE), allows the company to extract mechanistic immune insights and guide decision-making in drug development and clinical trials. IDE is composed of two AI models: AMICA ONE, a transformer-based model that pinpoints correlations of biological interest and makes predictions, such as identifying toxicity or optimal dosing, and AMICA REASON, a module that provides explainability of insights, which Solomon describes as an area of high investment for Immunai in 2025 and 2026.
“We believe it’s not enough to get good predictions. Explainability in biology, and in immunology in particular, is complicated,” said Solomon. “When you have an insight that you can explain, that’s when people are going to listen and follow the recommendations.”
Taken together, Solomon views the PICI collaboration as an opportunity to position AMICA as a valuable tool for Immunai’s partners.
“I hope that by sharing what we can find with AMICA ONE and AMICA REASON on the RADIOHEAD dataset, the community will get a glimpse of the platform and work with us even further,” said Solomon.
The post Immunai and Parker Institute to Build Massive Cancer Single-Cell Dataset appeared first on GEN - Genetic Engineering and Biotechnology News.