RESEARCH
Embedded Files
CAR-VISTA
CAR-VISTA
CAR-VISTA
CAR-VISTA
CAR-VISTA
CAR-VISTA
CAR-VISTA is an ongoing open-source project developing a hybrid AI–mechanistic simulation framework for virtual testing of CAR-T and CAR-NK immunotherapy strategies. The platform is designed to integrate mechanistic immune–tumor dynamics, spatial tissue structure, omics-informed cell states, and machine-learning prediction to study how engineered immune cells interact with tumors and the surrounding microenvironment.
The goal is to move beyond static response prediction toward counterfactual, intervention-aware modeling. CAR-VISTA will allow researchers to simulate alternative cell designs, dosing strategies, tumor phenotypes, resistance mechanisms, and microenvironmental constraints in silico before experimental validation, supporting more interpretable and efficient immunotherapy optimization.
The goal is to move beyond static response prediction toward counterfactual, intervention-aware modeling. CAR-VISTA will allow researchers to simulate alternative cell designs, dosing strategies, tumor phenotypes, resistance mechanisms, and microenvironmental constraints in silico before experimental validation, supporting more interpretable and efficient immunotherapy optimization.
SenesceNet
SenesceNet
CytoHybrid-OS
CytoHybrid-OS
SenesceNet
SenesceNet
CytoHybrid-OS is an open-source virtual tissue modeling framework that connects biological data to dynamic, interpretable simulation. The platform takes inputs such as single-cell RNA sequencing, spatial transcriptomics, proteomics, imaging, and perturbation data, then uses them to initialize virtual tissues containing immune and stromal cells with flexible internal gene-program states. Instead of treating cells as fixed labels, CytoHybrid-OS models functional identity as an emergent property shaped by inflammation, cytotoxicity, exhaustion, immune regulation, repair, fibrosis, matrix remodeling, local signals, and cell-cell communication.
The framework is designed to support counterfactual biological testing: researchers can simulate interventions, generate RNA-like and spatial readouts, compare mechanistic and AI/ML models, and evaluate whether predictions remain reliable after treatment changes the underlying biology. Its benchmark system compares pure machine learning, mechanistic models, hybrid mechanis
The framework is designed to support counterfactual biological testing: researchers can simulate interventions, generate RNA-like and spatial readouts, compare mechanistic and AI/ML models, and evaluate whether predictions remain reliable after treatment changes the underlying biology. Its benchmark system compares pure machine learning, mechanistic models, hybrid mechanis
Page updated
Google Sites
Report abuse