Introducing the novel Cytoscape app TimeNexus to analyze time-series data using temporal MultiLayer Networks (tMLNs)

Michaël Pierrelée,Bianca H Habermann,Laurent Tichit,Ana Reynders,Fabrice Lopez,Aziz Moqrich

doi:10.1038/s41598-021-93128-5

Abstract

Integrating -omics data with biological networks such as protein–protein interaction networks is a popular and useful approach to interpret expression changes of genes in changing conditions, and to identify relevant cellular pathways, active subnetworks or network communities. Yet, most -omics data integration tools are restricted to static networks and therefore cannot easily be used for analyzing time-series data. Determining regulations or exploring the network structure over time requires time-dependent networks which incorporate time as one component in their structure. Here, we present a method to project time-series data on sequential layers of a multilayer network, thus creating a temporal multilayer network (tMLN). We implemented this method as a Cytoscape app we named TimeNexus. TimeNexus allows to easily create, manage and visualize temporal multilayer networks starting from a combination of node and edge tables carrying the information on the temporal network structure. To allow further analysis of the tMLN, TimeNexus creates and passes on regular Cytoscape networks in form of static versions of the tMLN in three different ways: (i) over the entire set of layers, (ii) over two consecutive layers at a time, (iii) or on one single layer at a time. We combined TimeNexus with the Cytoscape apps PathLinker and AnatApp/ANAT to extract active subnetworks from tMLNs. To test the usability of our app, we applied TimeNexus together with PathLinker or ANAT on temporal expression data of the yeast cell cycle and were able to identify active subnetworks relevant for different cell cycle phases. We furthermore used TimeNexus on our own temporal expression data from a mouse pain assay inducing hindpaw inflammation and detected active subnetworks relevant for an inflammatory response to injury, including immune response, cell stress response and regulation of apoptosis. TimeNexus is freely available from the Cytoscape app store at https://apps.cytoscape.org/apps/TimeNexus.

Highlights

Time-series gene or protein expression data can give invaluable insight into the temporal dynamics of biological processes
TimeNexus was developed with the idea to create a versatile framework for working with temporal multilayer networks in the Cytoscape environment (Fig. 2)
We introduced here TimeNexus, a Cytoscape app to create, manage and visualize temporal multilayer networks

Summary

Introduction

Time-series gene or protein expression data can give invaluable insight into the temporal dynamics of biological processes. Methods and protocols exist to analyze time-series expression data and extract the dynamically expressed genes from a temporal dataset, some of which have been reviewed and compared in[1] Results from such tools do not provide insights into the activity of key molecules or pathways at a given time point. The Cytoscape app D yNetViewer[31] is able to construct, analyze and visualize active temporal networks It provides four different algorithms for constructing one static active subnetwork for each time point by retaining only the active nodes from a large protein interaction network at that time point. It provides in addition network analysis functions, mostly focusing on centrality measures and graph clustering algorithms of the temporal network. TimeNexus can be used to generate, manage and visualize tMLNs

Methods

Results

Conclusion