Exploring the dynamics of large-scale gene regulatory networks using hardware acceleration on a heterogeneous CPU-FPGA platform

Abstract

Systems Biology pose computational challenges that surpass the capabilities of current computing platforms. A particular challenge in this regard is simulating the dynamics of Gene Regulatory Networks (GRNs). This is because the number of possible network states, and thus the time required to compute them, grows exponentially with the number of network components. FPGA-based accelerators appear as promising alternatives to simulations on conventional software platforms. However, their application in the dynamic simulation of GRNs has been hindered by the inaccessibility of this technology to users without the proper expertise. Heterogeneous CPU-FPGA computing platforms, combined with high-level, flexible tools to deploy FPGA implementations, represent a powerful avenue to open the benefits of hardware acceleration to biologists. We present a high-level framework that exploits such acceleration for the simulation of GRNs, without compromising flexibility. Our tool takes as an input a biological network description in a standard format and translates it into a Verilog module. This module is subsequently incorporated into a hardware design optimized for the simulation of the network dynamics. To demonstrate the potential of our methodology, we simulated the dynamics of two large-scale GRNs, one involved in tumor formation in Colitis-Associated Colon cancer consisting of 70 nodes, and the other describing the signaling pathways of a lymphocyte cell comprised of 188 nodes. Our approach simulates the dynamics of the networks up to two orders of magnitude faster than an optimized OpenMP implementation on a multicore CPU.