A cell size- and cell cycle-aware stochastic model for predicting time-dynamic gene network activity in individual cells.

Ruijie Song,Ping Liu,Murat Acar,Weilin Peng

doi:10.1186/s12918-015-0240-5

Abstract

BackgroundDespite the development of various modeling approaches to predict gene network activity, a time dynamic stochastic model taking into account real-time changes in cell volume and cell cycle stages is still missing.ResultsHere we present a stochastic single-cell model that can be applied to any eukaryotic gene network with any number of components. The model tracks changes in cell volume, DNA replication, and cell division, and dynamically adjusts rates of stochastic reactions based on this information. By tracking cell division, the model can maintain cell lineage information, allowing the researcher to trace the descendants of any single cell and therefore study cell lineage effects. To test the predictive power of our model, we applied it to the canonical galactose network of the yeast Saccharomyces cerevisiae. Using a minimal set of free parameters and across several galactose induction conditions, the model effectively captured several details of the experimentally-obtained single-cell network activity levels as well as phenotypic switching rates.ConclusionOur model can readily be customized to model any gene network in any of the commonly used cells types, offering a novel and user-friendly stochastic modeling capability to the systems biology field.Electronic supplementary materialThe online version of this article (doi:10.1186/s12918-015-0240-5) contains supplementary material, which is available to authorized users.

Highlights

Despite the development of various modeling approaches to predict gene network activity, a time dynamic stochastic model taking into account real-time changes in cell volume and cell cycle stages is still missing
Modeling cell volume growth and division Our stochastic single-cell model consists of two interrelated modules
In this paper, we present a single-cell level stochastic model that accounts for the cell volume and the cell cycle in addition to the gene network it models, and demonstrate its predictive power by using the GAL network in S. cerevisiae

Summary

Introduction

Despite the development of various modeling approaches to predict gene network activity, a time dynamic stochastic model taking into account real-time changes in cell volume and cell cycle stages is still missing. Much work has been done to computationally model gene expression networks, including the well-characterized galactose utilization network (GAL network) in yeast Many of those models [6,7,8,9], are deterministic models and could provide only limited insights on what happens at the single-cell level. The shortcomings of this approach is demonstrated by previous work [10] that showed that stochastic noise could generate bimodality in a system whose deterministic models predict no bistability. Such approximations could be acceptable if the simulations lasted for a time period much shorter than the duration of the cell cycle, but they would be questionable for longer simulation durations

Methods

Results

Discussion

Conclusion