Parse Biosciences, the single cell sequencing company owned by QIAGEN, and proteomics company Nomic Bio are building a joint foundational dataset for immune biology called Cascade. The dataset will combine 100 million single cell RNA sequencing profiles with more than 11,000 proteomic secretome profiles, capturing how immune cells respond across more than 150,000 individual and combinatorial cytokine and drug perturbation conditions at multiple timepoints.
The design is deliberately multimodal and time resolved. Samples come from peripheral blood cells of 12 healthy donors, covering at least 14 immune cell types. Parse GigaLab will generate the single cell data, and Nomic Omni 1000 platform will measure more than 1,000 secreted proteins from the same samples, so that cell state changes and the protein signals they produce can be modelled together rather than inferred from separate experiments.
The commercial signal is the partner list. Cascade already has founding partner commitments from GSK, Pfizer and Sanofi, plus Graph Therapeutics and Turbine, and the dataset will be available for licensing, with generation underway. Three of the world largest drugmakers paying into a shared immune response reference is evidence that the constraint on AI driven immunology work is currently standardised, large scale perturbation data, not model access.
The claim to test is scale. The companies call Cascade the largest AI ready resource of its kind for immune modelling. Whether a healthy donor reference predicts responses in disease tissue, and whether licensees can turn it into candidate selection decisions that beat programme specific data, will only be visible once the data are delivered and used. For now Cascade is a data infrastructure deal, not a drug result, and it belongs in the same category as the lab automation financing elsewhere in this edition: capital moving into the inputs of AI drug discovery rather than into individual molecules.