Kozo Nishida - Automating biological network visualization with Jupyter Notebook | JupyterCon 2020

Brief Summary Cytoscape is the de facto standard software for biological network visualization. Cytoscape is mainly used as stand-alone desktop software. On the other hand, it has a REST server function and can be controlled with HTTP requests. In this talk, I will introduce how to use Cytoscape and Jupyter Notebook together as a biological visualization workflow automation system. Outline Introductions to Cytoscape and biological network visualization What is Cytoscape? Cytoscape is a highly customizable network visualization software. The styling feature according to the values assigned to nodes and edges is particularly useful to create figure for research paper. Biological network visualization It is a network in which molecules are represented by nodes and interactions between molecules are represented by edges. The nodes and edges have biological quantitative values. In the visualization workflow, various software and data are used to give the quantitative values to the network. An implementation of such a workflow for Cytoscape is called an “App”. Creating a Cytoscape App App has been written in Java mainly for GUI operation. Motivations for the automation of the biological network visualization workflow Why automation? GUI operation is manual and time consuming. Cytoscape has REST service and can be controlled with HTTP requests. By using REST service, the workflow can be automated with Python or R instead of Java. Automation of the operation improves the research reproducibility. Why Jupyter Notebook? Writing workflows in Jupyter Notebook improves their reproducibility. With Binder or Google Colab, anyone can learn the workflow interactively and check the reproducibility. The workflow system that combines Jupyter Notebook and Cytoscape RCy3 and py4cytoscape: REST service wrapper library for compatibility with Python and R packages Demonstrations Central thesis Biological network visualization workflow can be automated and reproducible with Jupyter Notebook