A parallelized cellular Potts model that enables simulations at tissue scale

Abstract

The Cellular Potts Model (CPM) is a widely used simulation paradigm for systems of interacting cells that has been used to study scenarios ranging from plant development to morphogenesis, tumour growth and cell migration. Despite their wide use, CPM simulations are considered too computationally intensive for three-dimensional (3D) models at organ scale. CPMs have been difficult to parallelise because of their inherently sequential update scheme. Here, we present a Graphical Processing Unit (GPU)-based parallelisation scheme that preserves local update statistics and is up to 3-4 orders of magnitude faster than serial implementations. We show several examples where our scheme preserves simulation behaviors that are drastically altered by existing parallelisation methods. We use our framework to construct tissue-scale models of liver and lymph node environments containing millions of cells that are directly based on microscopy-imaged tissue structures. Thus, our GPU-based CPM framework enables in silico studies of multicellular systems of unprecedented scale.