UQ-PhysiCell Documentation

Welcome to the UQ-PhysiCell documentation! This project provides uncertainty quantification tools for PhysiCell models.

Release: v1.2.8 (GitHub link)

Note

This readthedocs documentation is under active development.

Contents:

• About UQ-PhysiCell
  • Quick Start
• Installation
  • Requirements
  • Installation from PyPI
  • Installation from Source
  • Additional Dependencies
  • Verification
• Graphical User Interface (GUI)
  • Example Workflow
    • Step 1: Modify the Template Project Using PhysiCell Studio
    • Step 2: Create the UQ PhysiCell Configuration File (.ini)
    • Step 3: Generate the Simulation Database (.db)
    • Step 4: Define Quantities of Interest (QoIs) and Perform Sensitivity Analysis
• Model Analysis
  • Overview
  • Sensitivity Analysis
    • Global Sensitivity Analysis (GSA)
    • Local Sensitivity Analysis (LSA)
• Model Calibration
  • Bayesian Optimization
    • 1. Parameter Space (\(\Theta\))
    • 2. Quantities of Interest (QoIs)
    • 3. Distance Metrics
    • 4. Fitness Functions
  • Approximate Bayesian Computation (ABC)
  • References
• API Reference
  • Core UQ-PhysiCell Module
    • PhysiCell_Model
  • Model Analysis Module
    • ModelAnalysisContext
    • run_simulations()
    • Sampling Methods
    • Sensitivity Analysis
    • Utils
  • Bayesian Optimization Module
    • CalibrationContext
    • single_objective_bayesian_optimization()
    • multi_objective_bayesian_optimization()
    • run_bayesian_optimization()
    • Plotting and Visualization
  • Approximate Bayesian Computation Module
    • CalibrationContext
    • run_abc_calibration()
  • Database
    • Model Analysis
    • Bayesian Optimization
  • Utils
    • Summary of simulation
    • Wrapper of output
    • Distances
• Examples
  • Example 1: Context Setup and Internals — Educational Walkthrough
    • Step 1 — Build the context and inspect its internals
    • Step 2 — Provide a sample and run
  • Example 2: Data Storage Modes — Precomputed QoIs vs Raw MCDS
    • Mode A — Precomputed QoIs
    • Mode B — Raw MCDS Storage
    • Querying both databases with the same call
  • Example 3: Global Sensitivity Analysis — Sobol with Multi-task Parallelization
    • Generate samples and run simulations
    • Summary statistics across replicates
    • Sobol sensitivity indices (S1 and ST)
  • Example 4: Global Sensitivity Analysis — Sobol with MPI
    • Run simulations via MPI
    • Load results and compute summary statistics
    • Sobol sensitivity indices (S1 and ST)
  • Example 5: Local Sensitivity Analysis — One-At-a-Time (OAT) Sampling
    • Generate samples and run simulations
    • Run simulations
    • Query any QoI post-hoc from the stored MCDS database
  • Example 6: Global Parameter Sampling — Latin Hypercube Sampling (LHS)
    • Generate samples and run simulations
    • Run simulations
    • Query any QoI post-hoc from the stored MCDS database
  • Example 7: Model Calibration — Bayesian Optimization
    • Create context and run Bayesian Optimization
    • Pareto front analysis
    • Visualize QoIs from Pareto front points
    • Visualize QoI fitness vs parameter values
  • Example 8: Model Calibration — Approximate Bayesian Computation (ABC-SMC)
    • Distance functions, prior, and ABC options
    • Create context and run ABC-SMC calibration
    • Posterior analysis
  • Example 9: Model Calibration — Bayesian Optimization with Warm Start
    • Pareto front analysis
    • Visualize QoIs from Pareto front points
    • Visualize QoI fitness vs parameter values
  • Example 10: Rule Inactivation Sensitivity Analysis — Pancreatic Cancer Invasion Model
    • Map sample IDs to rules and initial conditions
    • Compute RMSD relative to the nominal (all-rules-active) model
    • Normalize and plot RMSD per rule and initial condition
    • Rank rules by mean normalized RMSD across all QoIs
    • Cell snapshots — nominal vs top-sensitivity inactive rules
• Publications
• Funding