NFsim: A Novel Agent-based Platform for Stochastic Simulation of Complex Biological Systems

NFsim: A Novel, Agent-based Platform for Stochastic Simulation of Complex Biological Systems

Michael W. Sneddon1,3, James R. Faeder2, Thierry Emonet1,3

1) Interdepartmental Program in Computational Biology & Bioinformatics, Yale University

2) Department of Computational Biology, University of Pittsburgh School of Medicine

3) Department of Molecular, Cellular, and Developmental Biology, Yale University

During many biological and cellular processes, the formation of large molecular complexes, aggregates, and polymer chains lead to a combinatorial explosion in the number of possible molecular configurations and interactions. Standard biochemical modeling approaches require the enumeration of the complete state-space of a system, often making detailed biochemical simulations infeasible. To meet this challenge, we have developed NFsim, an extensible agent-based platform for efficiently simulating biochemical reaction systems that exhibit a high degree of combinatorial complexity. Individual molecules are treated as agents that are connected through bonds, so that only the molecular configurations that exist need to be accounted for. Reaction rules specified in the BioNetGen language define how molecular agents are created, destroyed, and interact. To propagate the system, NFsim implements a stochastic event scheduler based on a recent extension to the Gillespie algorithm. NFsim also includes the ability to define reaction rates as functional expressions of an agent’s local context, which enables powerful and efficient coarse-graining of complex biochemical processes. Finally, the NFsim architecture allows biochemical simulations to be encapsulated in autonomous agents, so that higher order cellular interactions and behavior can be modeled. We anticipate NFsim will have practical applications to many complex biological systems. Here, we demonstrate the power and flexibility of NFsim by constructing a stochastic model of chemotactic bacteria that navigate in a realistic, time evolving, 3D environment.

Contact: michael.sneddon@yale.edu

• Presentation in PDF