Interactive Multiresolution Visualization of Cellular Network Processes

Abstract

Computational models of network-driven processes have become a standard to explain cellular systems-level behavior and predict cellular responses to perturbations. Modern models can span a broad range of biochemical reactions and species that, in principle, comprise the complexity of dynamic cellular processes. Visualization plays a central role in the analysis of biochemical network processes to identify patterns that arise from model dynamics and perform model exploratory analysis. However, most existing visualization tools are limited in their capabilities to facilitate mechanism exploration of large, dynamic, and complex models. Here, we present PyViPR, a visualization tool that provides researchers static and dynamic representations of biochemical network processes within a Python-based Literate Programming environment. PyViPR embeds network visualizations on Jupyter notebooks, thus facilitating integration with Python modeling, simulation, and analysis workflows. To present the capabilities of PyViPR, we explore execution mechanisms of extrinsic apoptosis in HeLa cells. We show how community-detection algorithms can identify groups of molecular species that represent key biological regulatory functions and simplify the apoptosis network by placing those groups into interactively collapsible nodes. We then show how dynamic execution of a signal, under different kinetic parameter sets that fit the experimental data equally well, exhibit significantly different signal-execution modes in mitochondrial outer-membrane permeabilization -the point of no return in extrinsic apoptosis execution. Therefore, PyViPR aids the conceptual understanding of dynamic network processes and accelerates hypothesis generation for further testing and validation.1 Cellular processes are controlled by networks of biomolecular interactions that process 2 signals and trigger a response [1-3]. These molecular networks give rise to nonlinear 3 dynamic processes that are difficult to explain and predict using reductionist 4 methods [4]. Mathematical models of cellular signaling pathways have become 5 commonplace in order to gain insights and describe the molecular mechanisms that 6 control cellular processes [5][6][7]. In general, these models continue to grow in size and 7 complexity, which makes the exploration of network structure and dynamics increasingly 8 challenging. Visualization tools comprise one effective way to explore network processes 9 and acquire conceptual insights about signal-execution mechanisms. In addition, 10 visualization tools can facilitate detection of execution patterns, and aid in hypothesis 11 generation for experimental validation. However, to the best of our knowledge, most 12 tools focus on static network representations of models without strategies to deal with 13 June 3, 2019 1/17 increasingly complex models, and generally lack support to visualize model dynamics. 14 Therefore, there is a need for novel tools that provide viable visualizations of large 15 models as well as support for intuitive visualiza...