Abstract 4921: A new method for identifying a data-consistent self-reconfigurable predictive bio-network model of the cell cycle based on time series data and its application in cancer systems biology

Abstract

Abstract A new method for (re)constructing interaction networks using a limited number of time series data has been developed. This method may provide better opportunities for immediate identification of data-consistent models of biological systems. It may serve as an alternative or complementary approach to other existing data-driven strategies for modeling or mining time-evolutionary properties of complex biological processes based on the analysis of their time series data. We present some practically useful data-mining techniques for constructing self-reconfigurable predictive models of complex biological systems, which approximate their underlying processes. These parametric models, if well reverse-engineered, may help capture key features of data relevant for the purpose of interest. To demonstrate our network inference method, we focus on analyzing time series data of cdk2, cyclinA1, cyclinD1, cycinD2, cyclinD3, E2F1, p16 and p27 obtained from real experiments. The constructed predictive model of the cell-cycle is found to be data-consistent. This model approximates the underlying biological processes hidden in the data, and may help reveal or identify key processes that may govern G1-S phase progression. Given multiple sets of time series data from a cell line, where some sets represent control conditions and other intervention conditions, the method introduced here can help construct interaction networks for each of these data sets. To investigate areas of the signaling network most affected by the intervention, critical areas that have been identified in those networks could be compared to the effects of real perturbations. This may help inform future experimental design by targeting sensitive areas in the signaling network or avoiding resistant pathways. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 4921. doi:1538-7445.AM2012-4921