Grid computing for sensitivity analysis of stochastic biological models

  • Authors:

  • Ivan Merelli;Dario Pescini;Ettore Mosca;Paolo Cazzaniga;Carlo Maj;Giancarlo Mauri;Luciano Milanesi

  • Affiliations:

  • Istituto Tecnologie Biomediche, Consiglio Nazionale Ricerche, Segrate, Italy;Università degli Studi di Milano-Bicocca, Dipartimento di Statistica, Milano, Italy;Istituto Tecnologie Biomediche, Consiglio Nazionale Ricerche, Segrate, Italy;Università degli Studi di Bergamo, Dipartimento di Scienze della Persona, Bergamo, Italy;Università degli Studi di Milano-Bicocca, Dipartimento di Informatica, Sistemistica e Comunicazione, Milano, Italy;Università degli Studi di Milano-Bicocca, Dipartimento di Informatica, Sistemistica e Comunicazione, Milano, Italy;Istituto Tecnologie Biomediche, Consiglio Nazionale Ricerche, Segrate, Italy

  • Venue:
  • PaCT'11 Proceedings of the 11th international conference on Parallel computing technologies
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

Systems biology is a multidisciplinary research area aimed at investigating biological systems by developing mathematical models that approach the study and the analysis of both the structure and behaviour of a biological phenomenon from a system perspective. The dynamics described by such mathematical models can be deeply affected by many parameters, and an extensive exploration of the parameters space in order to find crucial factors is most of the time prohibitive since it requires the execution of a huge number of computer simulations. Sensitivity analysis techniques can help in understanding how much the uncertainty in the model outcome is determined by the uncertainties, or by the variations, of the model input factors (components, reactions and respective parameters). In this work we exploit the European Grid Infrastructure to manage the calculations required to perform the SA on a stochastic model of bacterial chemotaxis, using an improved version of the first order screening method of Morris. According to the results achieved in our exploratory analysis, the European Grid Infrastructure is a useful solution for distributing the stochastic simulations required to carry out the SA of a stochastic model. Considering that the more intensive the computation the more scalable the infrastructure, grid computing can be a suitable technology for large scale biological models analysis.