Track genomics, neuroscience and molecular biology.
Search papers, labs, and topics across Lattice.
To address the fragmentation and unreliability of multi-model in silico biology, PRAXIS-VirtualCell establishes a modular runtime that unifies biological data, predictive models, perturbations, and execution environments across species including *E. coli*, yeast, and human K562 cells. The framework matters because existing virtual cell systems lack verifiable boundaries, often failing silently when generalizing outside their training data. By coupling agentic natural-language orchestration with formal biological contracts, the system autonomously translates experimental queries into reproducible pipelines while explicitly gating predictions through evidence-aware abstention.
Virtual cell models often hallucinate outside their training distribution, but anchoring agentic workflows to explicit biological contracts enables automated in silico experimentation with principled abstention across species.
Virtual cells are evolving from single-task predictive models toward programmable biological simulation systems, yet heterogeneous data, models, and validation evidence still lack a unified organizational framework. Here, we present PRAXIS-VirtualCell, a modular framework that organizes biological data, predictive models, perturbations, adapters, execution environments, and validation evidence to enable reproducible and auditable virtual experiments. The system supports cross-species tasks spanning Escherichia coli, Saccharomyces cerevisiae, and human K562 cells, while using biological contracts and evidence-aware execution to distinguish supported predictions from extrapolation and abstention. By further integrating agentic orchestration, PRAXIS-VirtualCell automatically translates natural-language questions into traceable virtual experiments, providing a unified runtime foundation for trustworthy and scalable Virtual Cell systems.