Abstract
Colorectal cancer (CRC) cell atlases have refined descriptive maps of tumour ecosystems, yet cross-sample integration often obscures disease-relevant patient-specific variation and remains largely correlative, limiting insight into the mechanisms and state transitions that drive progression and treatment response. Here, we develop a continual learning framework to construct a comparative single-cell CRC atlas spanning over 300 patients and 1.5 million cells, preserving inter-patient variation while aligning healthy and malignant contexts. We resolve distinct non-canonical malignant cell states, including an endoderm-like state enriched in microsatellite-stable, KRAS -mutant CRC with features of oncofetal plasticity. Cell states are recapitulated in patient-derived organoids, establishing a tractable model of reprogramming. By linking the observational atlas to a large-scale perturbation atlas using relative representations, we map perturbations that drive cells toward defined phenotypic extremes. We connect cell states to therapeutic responses, showing that MAPK inhibition induces a shift away from a proliferative phenotype and converges towards a plastic, endoderm-like state. Together, this framework moves beyond static atlases to enable mechanistic modeling of cell-state regulation and causal inference toward cell-state–directed therapies.
Full text
2,401 characters
· extracted from
oa-doi-fallback
· click to expand
Abstract
Colorectal cancer (CRC) cell atlases have refined descriptive maps of tumour ecosystems, yet cross-sample integration often obscures disease-relevant patient-specific variation and remains largely correlative, limiting insight into the mechanisms and state transitions that drive progression and treatment response. Here, we develop a continual learning framework to construct a comparative single-cell CRC atlas spanning over 300 patients and 1.5 million cells, preserving inter-patient variation while aligning healthy and malignant contexts. We resolve distinct non-canonical malignant cell states, including an endoderm-like state enriched in microsatellite-stable, KRAS-mutant CRC with features of oncofetal plasticity. Cell states are recapitulated in patient-derived organoids, establishing a tractable model of reprogramming. By linking the observational atlas to a large-scale perturbation atlas using relative representations, we map perturbations that drive cells toward defined phenotypic extremes. We connect cell states to therapeutic responses, showing that MAPK inhibition induces a shift away from a proliferative phenotype and converges towards a plastic, endoderm-like state. Together, this framework moves beyond static atlases to enable mechanistic modeling of cell-state regulation and causal inference toward cell-state–directed therapies.
Competing Interest Statement
R.E is a co-founder and holds equity in Ensocell Therapeutics. In the past three years, G.P. has been a consultant for Mosaic Therapeutics. S.A.T. is a scientific advisory board member of Bioptimus, ForeSite Labs, Xaira Therapeutics, Board observer and equity holder of TransitionBio, a co-founder, consultant and Board Director of Ensocell Therapeutics, a non-executive director of 10x Genomics and a part-time employee of GlaxoSmithKline. F.J.T. consults for Immunai, CytoReason, BioTuring and Phylo Inc., GenBio, and Valinor Industries, and has ownership interest in RN.AI Therapeutics, Dermagnostix, and Cellarity. AstraZeneca, GlaxoSmithKline, and Astex Pharmaceuticals have awarded M.J.G. research grants. M.J.G. is a consultant for Bristol Myers Squibb. M.J.G. is a board director for and equity holder in Mosaic Therapeutics. The remaining authors declare no competing interests.
Footnotes
Three of the references were misplaced in version 2. A correction has been applied in the new version.
Text is read by the "Ask this paper" AI Q&A widget below.
Extraction quality varies by source — PMC NXML preserves structure
cleanly, OA-HTML may include some navigation residue, and OA-PDF can
have broken hyphenation. The publisher copy
(via DOI)
is the canonical version.