An open model of every cancer cell state, built from perturbation atlases
Map every state a cancer cell can be in, and how drugs and the surrounding tissue move it between states, into an open computational model anyone can query and improve.
Single-cell and spatial atlases (Human Tumor Atlas Network, Human Cell Atlas) describe cell states; perturbation screens and foundation models trained on them begin to predict responses. The proposal is a coordinated, openly licensed effort to generate perturbation-response single-cell data across hundreds of models and patient samples, train and release a foundation model of cancer cell state transitions, and benchmark it prospectively against drug response in organoids and trials, with the data, weights and benchmarks all public.
- Tumour heterogeneity and clonal evolution · A tumour is many tumours. Treatments that kill most cells leave the rest to grow back, changed.
- Preclinical models that do not predict people · Nine in ten cancer drugs that work in mice fail in humans. Our models are the reason.
- Preclinical results do not reproduce · Fewer than half of landmark cancer biology findings reproduce when someone else tries.
Pages like this
not linked directly; found by shared links- IdeaBarcode patient-derived tumours to watch which clones win under each drug
Shares CRISPR functional genomics, Patient-derived organoids, Tumour heterogeneity and clonal evolution, Preclinical models that do not predict people.
- IdeaScore every model system on how well it predicted real trial results
Shares Broad Institute of MIT and Harvard, Patient-derived organoids, Preclinical results do not reproduce, Preclinical models that do not predict people.
- IdeaSelf-driving laboratories that run the cancer biology hypothesis loop autonomously
Shares CRISPR functional genomics, Patient-derived organoids, Preclinical results do not reproduce, Preclinical models that do not predict people.
- IdeaMap which tumour clones sit next to which immune cells before choosing therapy
Shares 10x Genomics, Single-cell & spatial profiling, Tumour heterogeneity and clonal evolution.
- InstitutionThe Jackson Laboratory Cancer Center
Shares CRISPR functional genomics, Preclinical results do not reproduce, Preclinical models that do not predict people.
- IdeaA bone marrow niche on a chip to study human dormancy
Shares Single-cell & spatial profiling, Patient-derived organoids, Preclinical models that do not predict people.
- IdeaTest drugs on freshly cut slices of the patient's own tumour
Shares Single-cell & spatial profiling, Patient-derived organoids, Preclinical models that do not predict people.
- IdeaAn open atlas of collateral sensitivity for every approved targeted drug
Shares Broad Institute of MIT and Harvard, CRISPR functional genomics, Tumour heterogeneity and clonal evolution.