Blog Post

Cytoscape and customizing your network

Cytoscape and customizing your network

"Cytoscape is an open source and [Java-based] software for integrating biomolecular interaction networks with high-throughput expression data and other molecular states into a unified conceptual framework" (Shannon 2498).

It uses graph theory for modeling and rendering to display interactive graphs.

This application is very professional with a lot of features and requires good knowledge of Data analysis to be able to work with it. During the class we created a network visualization by modified an excel file (Source, Action and Target) and importing it into the application. We had to follow all of those instructions carefully.

There are two data sets we used them from, 2015 and 2016:

Fist of all we modified both datasaet (csv files) by keeping lawenforcementagency, classification and raceethnicity columns and then removing the other columns and then save every file to the xslx format.

Then, I could make many different visualizing by changing the style's properties. There are three tabs are associated with "Style": Node, Edge and Net. By changing the Style, you will have different visualization (color, shape...) 


In the bottom of the main panel, there are three tabs (Node Table, Edge Table and Network Table). 

Edge Table has four properties "shared name", "shared interaction", "name" and "description”. You can select one or some of them (Ctrl + Click) and then "apply to selected edges".

See different visualize after new changes (red lines)

This is another snapshot with captions:

There are tons of features and tools that you can modify your results. But it is not easy to figure out.

The following graph is based on 2016's data which I created with Cytoscape. The blue spots represent police stations and the red spots show ethnicities. Here we see that there is a significant difference between graphs of "Arab-Americans" and "White" (or "Black"). Since I did not have data of the population of ethnicities and their geographical locations, I could not come up with a precise analysis of the relationships between the ethnicities and shootings. Therefore, I thought that I can compare topologies of those graphs with each other. Here the simplest graph is related to the "Arab-American" which has only three edges. The graph "Native American" has more edges and has a connection edge with to the "White" graph and an intersection with the "Hispanic-Latino" graph.

The "White" graph has at most edges and has connections with all graphs except "Arab-American". "Hispanic-Latino" and "Black" graphs are more complicated. Because we can see there are some multiple edges between single blue node and single red nodes. For example, there are multiple edges between "Chicago Police Department" and "Black". More interesting that there are also multiple edges between "Chicago Police Department" and "White".


Works Cited:

Shannon, Paul, et al. "Cytoscape: a software environment for integrated models of biomolecular interaction networks." Genome research 13.11 (2003): 2498-2504, Accessed 8 March 2017.



No comments