PhysiCell

PhysiCell simulates physics-based, agent-based multicellular systems to model cellular behaviors and microenvironmental biochemical transport for studying tissue-scale processes such as tumor growth and immune interactions.

Key Features:

• Multi-Substrate Biotransport Solver: Links cell phenotypes with multiple diffusing substrates and signaling factors to simulate biochemical interactions within tissues.
• Biologically-Driven Sub-Models: Provides built-in models for cell cycling, apoptosis, necrosis, solid and fluid volume changes, mechanics, and motility.
• Performance and Scalability: Implemented in C++ with minimal dependencies and parallelized with OpenMP to scale to ~10^5–10^6 cells on desktop workstations and larger simulations on HPC nodes.
• Application Demonstrations: Supports simulations of necrotic core biomechanics and stochasticity in tumor spheroids and ductal carcinoma in situ (DCIS), and exploration of synthetic multicellular systems.
• Extensibility: Allows incorporation and modification of additional models to replicate other simulation platforms or extend research-specific behaviors.
• Integration for Model Exploration: Integrates with EMEWS to enable dynamical simulation on HPC and adaptive parameter-space exploration using active learning and genetic algorithms.

Scientific Applications:

• Tumor spheroids and DCIS: Simulation of necrotic core biomechanics and stochastic growth dynamics in tumor spheroids and ductal carcinoma in situ.
• Synthetic multicellular systems for therapy: Exploration of engineered multicellular systems for potential anti-cancer treatments.
• Cancer heterogeneity: Investigation of intratumoral heterogeneity and its impact on tumor dynamics.
• Cancer immunology and immunotherapy design: Modeling immune–tumor interactions and enabling adaptive exploration of immunotherapy design spaces via integration with EMEWS.

Methodology:

Physics‑based agent-based simulations with a multi-substrate biotransport solver linking diffusing substrates and signaling factors to cell phenotypes; built-in submodels for cell cycling, apoptosis, necrosis, volume changes, mechanics, and motility; implemented in C++ and parallelized with OpenMP; integration with EMEWS using active learning and genetic algorithms for adaptive parameter sampling on HPC.