Home

CNA-ResistDynamics is a research software package that infers the temporal evolution of chemotherapy resistance in ovarian cancer from longitudinal liquid-biopsy data.

It uses two complementary dynamical models:

• ODE model — well-mixed population dynamics of sensitive (S) and resistant (R) tumour cells under periodic cytotoxic treatment
• PDE model — 1-D reaction–diffusion system that adds spatial structure to the resistance dynamics

Resistance is measured via subclonal CNA fractions estimated by the liquidCNA algorithm, combined with CA125 serum protein as a tumour-burden proxy.

---

Key features

• Likelihood-based multi-start optimisation (L-BFGS-B for ODE, Powell for PDE)
• Numba-JIT hot paths for ODE simulation and negative log-likelihood evaluation
• Operator-splitting FEniCS/dolfinx 1-D PDE solver with PETSc system caching
• Optional PyMC Bayesian posteriors (SMC / Metropolis)
• Sobol sensitivity analysis via SALib
• 2-D FEniCS + PyVista mesh visualisation pipeline
• Unified tumorfits CLI covering every workflow step
• Snakemake workflow with config.yaml as the single source of truth

---

Quick navigation

• Quick Start — Run your first fit in minutes
• Mathematical Model — ODE and PDE equations
• Data — Data formats and preparation
• CLI Reference — All commands and options
• API Reference — Python package documentation

---

Citation

If you use this software, please cite:

Hockings et al. (2025). Subclonal copy-number alterations drive resistance to platinum-based chemotherapy in ovarian cancer. Cancer Research. DOI: 10.1158/0008-5472.CAN-25-0351