CellType | Biological Reasoning Engine

We simulate drug response in human biology so you can prioritize what to test and accelerate what works. We train foundation models on millions of human cell profiles so they learn the patterns that govern drug response, toxicity, patient variation, and diagnostic signals. Published at top ML venues. Featured by Google. Simulate drug response across targets so you commit screening budget to the right bets. Predict human‑specific toxicity and organ risk before IND — catching signals animal models miss. Simulate patient response so you design the right trial the first time. Assay‑ready biomarkers (gene panels, IHC, inclusion/exclusion rules) for go/no‑go decisions. Foundation models for multimodal biological data — integrating across genomics, methylation, proteomics, and more to extract signal that single-modality methods miss. C2S-Scale — a 27B-parameter model trained on 1B+ tokens of transcriptomic data. Results from C2S-Scale (2026). Evaluated on standard single-cell benchmarks. The

saved by

• Eva Seigneurbieux