A Problem Solving Environment for Stochastic Biological Simulations

Stochastic simulations of biological systems vary widely in scope from reaction modules, to single cells, to cell colonies. While the same techniques for sampling the stochastic equations governing cellular processes apply to all these systems, the setup of the simulation volume and initial state for the simulations differ significantly. Lattice Microbes is a GPU accelerated stochastic biological problem solving environment with a general interface that meets the diverse requirements of these types of biological simulations. The software includes a Python interface that allows facile customization of the simulation setup and on-the-fly modification of the simulation state with access to highly optimized, compiled algorithms for solving the stochastic equations. Here we describe the interface to Lattice Microbes and present several examples of very different simulations that were rapidly prototyped in Python. Two examples show standard stochastic biochemical problems. As an example of the true utility of the Python interface, the highly optimized method for sampling the reaction-diffusion master equation in Lattice Microbes is coupled to the COBRA toolbox, a Python package that solves a linear programming problem representing the steady-state reaction flux through the full metabolic model of the cell. This final example shows how the Python interface allows the GPU optimized code to be used to interface with other methods.