FiCoS: a fine-grained and coarse-grained GPU-powered deterministic simulator for biochemical networks

Abstract

Mathematical models of biochemical networks can largely facilitate the comprehension of the mechanisms at the basis of cellular processes, as well as the formulation of hypotheses that can then be tested with targeted laboratory experiments. However, two issues might hamper the achievement of fruitful outcomes. On the one hand, detailed mechanistic models can involve hundreds or thousands of molecular species and their intermediate complexes, as well as hundreds or thousands of chemical reactions, a situation generally occurring when rule-based models are analysed. On the other hand, the computational analysis of a model typically requires the execution of a large number of simulations for its calibration or to test the effect of perturbations. As a consequence, the computational capabilities of modern Central Processing Units can be easily overtaken, possibly making the modeling of biochemical networks a worthless or ineffective effort. To the aim of overcoming the limitations of the current state-of-the-art simulation approaches, we present in this paper FiCoS, a novel “black-box” deterministic simulator that effectively realizes both a fine- and a coarse-grained parallelization on Graphics Processing Units. In particular, FiCoS exploits two different integration methods, namely the Dormand–Prince and the Radau IIA, to efficiently solve both non-stiff and stiff systems of coupled Ordinary Differential Equations. We tested the performance of FiCoS against different deterministic simulators, by considering models of increasing size and by running analyses with increasing computational demands. FiCoS was able to dramatically speedup the computations up to 855 ×, showing to be a promising solution for the simulation and analysis of large-scale models of complex biological processes.Author summarySystems Biology is an interdisciplinary research area focusing on the integration of biology and in-silico simulation of mathematical models to unravel and predict the emergent behavior of complex biological systems. The ultimate goal is the understanding of the complex mechanisms at the basis of biological processes together with the formulation of novel hypotheses that can be then tested with laboratory experiments. In such a context, detailed mechanistic models can be used to describe biological networks. Unfortunately, these models can be characterized by hundreds or thousands of molecular species and chemical reactions, making their simulation unfeasible using classic simulators running on modern Central Processing Units (CPUs). In addition, a large number of simulations might be required to calibrate the models or to test the effect of perturbations. In order to overcome the limitations imposed by CPUs, Graphics Processing Units (GPUs) can be effectively used to accelerate the simulations of these detailed models. We thus designed and developed a novel GPU-based tool, called FiCoS, to speed-up the computational analyses typically required in Systems Biology.

SMGen: A generator of synthetic models of biochemical reaction networks

“…It is worth mentioning that obtaining realistic RBMs exhibiting non trivial dynamics is fundamental to perform in-depth computational analyses and comparisons among the existing and the novel simulators. Indeed, in the case of stable or flat dynamics, or when the overall behavior of the network is extremely fast and instantly exhausts all the reactants, the most advanced integration algorithms are able to simulate the emergent dynamics in just one computation step [20]. In such a case, the computational performance of the simulation tools is only partially assessed, thus hindering a fair comparison among the tools.…”

“…Unfortunately, these computational tasks require the execution of huge amounts of simulations, so that the capabilities of biochemical simulators running on Central Processing Units (CPUs) (see, e.g., [9–11]) can be easily overtaken. Thus, several simulators exploiting Graphics Processing Units (GPUs) have been lately introduced to reduce the running times (see, e.g., [12–20]).…”

“…The computational performance of several GPU-powered tools were assessed using randomly generated synthetic RBMs [14,19,20]. However, only a few generators of biochemical models have been proposed so far, hindering the possibility of having a common and well-defined benchmarking approach.…”

“…Since each ODE corresponds to a specific chemical species, a higher number of species generally lead to a higher parallelization, increasing the computational performance of 304 the simulator. On the contrary, considering that the number of the reactions composing 305 the biological system is roughly related to the length of each ODE, in terms of the 306 mathematical complexity, the higher the number of reactions the higher the number of 307 operations that must be performed by each thread, leading to a higher running time 308 [19,20].…”

“…The outcome of such analyses showed that some well-known and widely used 128 meta-heuristics, generally able to outperform all competitors in the optimization of 129 benchmark functions, obtained poor performance when used to solve the PE problem. 130 Moreover, SMGen was exploited to generate symmetric and asymmetric RBMs required 131 for the in-depth analyses and comparisons of the computational performance of different 132 biochemical simulators and to highlight their peculiarities [19]. In particular, these 133 analyses allowed for determining the best simulator to employ under specific conditions, 134 such as the size of the RBM and the total number of simulations to perform.…”