Large-scale immune response dataset partnership
Proteomics company Nomic Bio and Parse Biosciences, a provider of single-cell sequencing and a QIAGEN company, have partnered to generate Cascade, a large-scale multimodal dataset designed to establish a foundational reference for healthy human immune response.
Built using controlled perturbations across healthy donors, Cascade combines single-cell transcriptomic measurements with quantitative secretome proteomics to capture changes in immune cell states and secreted protein signals over time. The dataset will support the development and evaluation of AI models of immune biology, providing a framework that can be extended to disease-specific research.
Foundational partners already include biopharma companies GSK, Pfizer and Sanofi, alongside computational drug discovery companies Graph Therapeutics and Turbine.
Using peripheral blood mononuclear cells (PBMCs) from 12 healthy donors, the dataset will combine transcriptomic and quantitative secretome measurements across three timepoints. It will capture how at least 14 immune cell types respond to individual and combinatorial cytokine, chemokine and drug perturbations.
In total, Cascade will include 100 million single-cell RNA-seq profiles, more than 11,000 proteomic profiles and over 150,000 perturbation measurements.
Its time-resolved design aims to capture early signals, transient responses and delayed protein secretion, providing a more comprehensive picture of immune responses than single-timepoint measurements.
Parse’s GigaLab will perform the single-cell RNA sequencing, while Nomic’s Omni 1000 platform will profile more than 1,000 secreted proteins from the same samples. The quantitative protein measurements are designed to remain comparable across studies, allowing Cascade to be integrated with future datasets covering additional donors, perturbations and disease contexts.
Linking RNA and protein measurements by donor, perturbation and time will also allow researchers to develop and evaluate AI models, including testing how cellular responses predict downstream protein profiles and how models generalise to unseen biological data.
Proteomics for AI biology
“Many therapies aim to modulate cytokine signalling,” said Milad Dagher, Chief Executive Officer of Nomic Bio. “To model those responses with AI, we need a strong reference for how healthy human immune cells respond across perturbations and time, including the proteins they release. Cascade creates that foundation and a framework to build on for disease-specific questions.”
Charlie Roco, Co-founder and Chief Technology Officer of Parse Biosciences, added: “Cascade was designed from the beginning as a multimodal dataset. The scale and structure of the study give researchers a common experimental framework for modelling immune response today and extending those models as new biological contexts and datasets are added.”
Data generation is underway, with the dataset set to be made available for licensing.
Organisations interested in Cascade or working with Nomic on proteomic and multimodal datasets for AI and computational biology can find out more here.
