Identification of Critical Molecular Components in a Multiscale Cancer Model Based on the Integration of Monte Carlo, Resampling, and ANOVA

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.

Development of a diffusion-based mathematical model for predicting chemotherapy effects

“…For each parameter, we created 11 variations, through a range of +/−50% of the parameter's standard value and with a 10.0% variation interval. Note that a 50% variation has been assumed as reasonable in systems modeling analysis [14–16]. Accordingly, this generated a total of (11 × 11 =) 121 parameter variation pairs, covering a wider range of parameter space.…”

Understanding Drug Resistance in Breast Cancer with Mathematical Oncology

“…These efforts have produced encouraging results in understanding cancer drug resistance, as seen in other research areas such as identification of novel therapeutic targets [76–78], development of alternative therapeutic strategies [79], and prediction of surgical volume and tumor size [80]. Eventually, these data-driven models could help improve patient outcomes and reduce costs of cancer treatment.…”

Development of a sampling‐based global sensitivity analysis workflow for multiscale computational cancer models

“…Hence, we use a 'random sampling' of input parameters to render the large number of variation combinations computationally manageable. Specifically, as implemented in [9], we use the Latin hypercube sampling (LHS) method to generate 2000 random sets of parameter values, and thus 2000 sets of simulation results will be generated correspondingly. For simplicity, we call each set of parameter values along with the corresponding two tumour output values an 'observation'.…”

“…GSA methods have also been applied to systems biology models [4,[6][7][8], but most of them focus on the analysis of signalling pathways. To assess the context-dependent relationship between different biological scales of interest, we have previously provided an applicable GSA strategy based on the integration of Monte Carlo and resampling methods as well as the repeated use of analysis of variance (ANOVA) [9]. Read et al [10] also developed a GSA method based on statistical techniques to link simulation results back into the original biology domain in order to determine the confidence of the simulation-derived predictions.…”