Identification of critical molecular components in a multiscale cancer model based on the integration of Monte Carlo, resampling, and ANOVA

Zhihui Wang, Veronika Bordas, Thomas S. Deisboeck

Research output: Contribution to journalArticle

17 Scopus citations

Abstract

To date, parameters defining biological properties in multiscale disease models are commonly obtained from a variety of sources. It is thus important to examine the influence of parameter perturbations on system behavior, rather than to limit the model to a specific set of parameters. Such sensitivity analysis can be used to investigate how changes in input parameters affect model outputs. However, multiscale cancer models require special attention because they generally take longer to run than does a series of signaling pathway analysis tasks. In this article, we propose a global sensitivity analysis method based on the integration of Monte Carlo, resampling, and analysis of variance. This method provides solutions to (1) how to render the large number of parameter variation combinations computationally manageable, and (2) how to effectively quantify the sampling distribution of the sensitivity index to address the inherent computational intensity issue. We exemplify the feasibility of this method using a two-dimensional molecular-microscopic agent-based model previously developed for simulating non-small cell lung cancer; in this model, an epidermal growth factor (EGF)-induced, EGF receptor-mediated signaling pathway was implemented at the molecular level. Here, the cross-scale effects of molecular parameters on two tumor growth evaluation measures, i.e., tumor volume and expansion rate, at the microscopic level are assessed. Analysis finds that ERK, a downstream molecule of the EGF receptor signaling pathway, has the most important impact on regulating both measures. The potential to apply this method to therapeutic target discovery is discussed.

Original languageEnglish (US)
Article numberArticle 35
JournalFrontiers in Physiology
VolumeJUL
DOIs
StatePublished - Dec 1 2011

Keywords

  • Agent-based model
  • Analysis of variance
  • Multiscale
  • Non-small cell lung cancer
  • Sensitivity analysis

ASJC Scopus subject areas

  • Physiology
  • Physiology (medical)

Fingerprint Dive into the research topics of 'Identification of critical molecular components in a multiscale cancer model based on the integration of Monte Carlo, resampling, and ANOVA'. Together they form a unique fingerprint.

Cite this