Search

The ErbB signalling pathway has been studied extensively owing to its role in normal physiology and its dysregulation in cancer. Reverse engineering by mathematical models use the reductionist approach to characterize the network components. For an emergent, system-level view of the network, we propose a data analytics pipeline that can learn from the data generated by reverse engineering and use it to re-engineer the system with an agent-based approach. Data from a kinetic model that estimates the parameters by fitting to experiments on cell lines, were encoded into rules, for the interactions of the molecular species (agents) involved in biochemical reactions. The agent model, a digital representation of the cell line system, tracks the activation of ErbB1-3 receptors on binding with ligands, resulting in their dimerization, phosphorylation, trafficking and stimulation of downstream signalling through P13-Akt and Erk pathways. The analytics pipeline has been used to mechanistically link HER expression profile to receptor dimerization and activation of downstream signalling pathways. When applied to drug studies, the efficacy of a drug can be investigated in silico. The anti-tumour activity of Pertuzumab, a monoclonal antibody that inhibits HER2 dimerization, was simulated by blocking 80% of the cellular HER2 available, to observe the effect on signal activation.

Motivation: Experiments in systems biology are generally supported by a computational model which quantitatively estimates the parameters of the system by finding the best fit to the experiment. Mathematical models have proved to be successful in reverse engineering the system. The data generated is interpreted to understand the dynamics of the underlying phenomena. The question we have sought to answer is that - is it possible to use an agent-based approach to re-engineer a biological process, making use of the available knowledge from experimental and modelling efforts? Can the bottom-up approach benefit from the top-down exercise so as to create an integrated modelling formalism for systems biology? We propose a modelling pipeline that learns from the data given by reverse engineering, and uses it for re-engineering the system, to carry out in-silico experiments., Results: A mathematical model that quantitatively predicts co-expression of EGFR-HER2 receptors in activation and trafficking has been taken for this study. The pipeline architecture takes cues from the population model that gives the rates of biochemical reactions, to formulate knowledge-based rules for the particle model. Agent-based simulations using these rules, support the existing facts on EGFR-HER2 dynamics. We conclude that, re-engineering models, built using the results of reverse engineering, opens up the possibility of harnessing the power pack of data which now lies scattered in literature. Virtual experiments could then become more realistic when empowered with the findings of empirical cell biology and modelling studies., Availability and Implementation: Implemented on the Agent Modelling Framework developed in-house. C ++ code templates available in Supplementary material ., Contact: liz.csir@gmail.com., Supplementary Information: Supplementary data are available at Bioinformatics online., (© The Author 2016. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: journals.permissions@oup.com)