CytoTrace2: Methods and Systems for Determining Phenotypic States from Genomic Data with Interpretable AI

Researchers at Stanford have developed a novel deep-learning-based tool called CytoTRACE2 that interprets single-cell RNA sequencing (scRNA-seq) to enable the discovery of regenerative cells across all tissue types and novel targets in cancer and other diseases.

CytoTRACE2 has been validated across 31 human and mouse scRNA-seq datasets encompassing 28 tissue types, outperforming existing methods in recapitulating experimentally determined potency levels and differentiation states. Moreover, it reconstructed the temporal hierarchy of mouse embryogenesis across 62 timepoints and facilitated discovery of cellular phenotypes in cancer linked to survival and immunotherapy resistance.

This technology has broad applicability for drug and biomarker discovery for diseases where developmental phenotypes and hierarchies play a role, especially cancer. It is also a powerful research tool for understanding the biology of these phenotypically complex diseases.

Stage of Development
Prototype