Contents

1 Version Information

R version: 4.1.0

Bioconductor version: 3.13

Cytoscape: 3.8.2

Cy3D (Cytoscape app): 1.1.3

KEGGscape (Cytoscape app): 0.9.1

2 Introduction

Visualization of Trans-omic networks helps biological interpretation by illustrating pathways where the signals are transmitted (Gehlenborg et al., 2010).

To characterize signals that go across multiple omic layers, Yugi and colleagues have proposed a method for network visualization (Yugi et al., 2014) by stacking multiple 2D pathways in a 3D space.

The 3D network visualization was realized by VANTED (Rohn et al., 2012). However, the visualization relies on time-consuming manual operation. Here we propose transomics2cytoscape, an R package that automatically creates 3D network visualization in combination with Cytoscape (Shannon, 2003), Cy3D App, and Cytoscape Automation (Otasek et al., 2019).

2.1 Installation

This package requires Cytoscape to be installed and you need to run Cytoscape before running the following R code.

BiocManager::install("transomics2cytoscape")

3 Workflow

There is 2 functions create3Dnetwork and createTransomicEdges in transomics2cytoscape. Below is a diagram of the transomics2cytoscape workflow.

Figure 1. The transomics2cytoscape workflow

Figure 1. The transomics2cytoscape workflow

create3Dnetwork has 3 arguments.

The 1st one is a directory path where you put the network files to be layered in 3D space. The 2nd one is a file path of TSV for the Z-axis layout of the network files (called “Layer definition file”). The last one is a file path of XML used to style Cytoscape.

For example,

suid <- create3Dnetwork(networkDataDir, networkLayers, stylexml)

createTransomicEdges has 2 arguments.

The 1st one is the SUID of the network created by create3Dnetwork. The 2nd one is a file path of TSV for the transomic interactions (called “Transomic interaction file”).

For example,

suid <- createTransomicEdges(suid, layer1to2)

3.1 Input files

3.1.1 (Any number of) network files to be layered in 3D space

Files that Cytoscape can import. You need to put these files in the directory of the 1st argument of create3Dnetwork. You don’t need to put files for the KEGG pathway. For KEGG pathway, you can import the network just by writing the KEGG pathway ID in the “Layer definition file” described later.

3.1.2 Layer definition file

“Layer definition file” is a TSV file for the Z-axis layout of the network files.

A file that defines network layer index and the Z-height of the network in 3D space. The format is as follows.

layer1  rno04910    2400
layer2  rno01100    1500
layer3  rno01100    1

The 1st column is the network layer index. This information is added to the node table column LAYER_INDEX.

The 2nd column is the KEGG pathway ID or the network file name in the directory of the 1st argument of create3Dnetwork. You don’t need to prepare a network file for the KEGG pathway. You can import the KEGG pathway simply by writing the KEGG pathway ID. The last column is the Z-height of the network.

3.1.3 A style file of Cytoscape

A Cytoscape style file. For more information about Cytoscape style file, see the Cytoscape user manual. Note that you can only use style properties that are supported by Cy3D.

3.1.4 Trans-omic interaction file

“Trans-omic interaction file” is a TSV file that defines the edges that connect the different network layers. The format is as follows.

Figure 2. Trans-omic interaction file

Figure 2. Trans-omic interaction file

The 1st ~ 4th columns are the information about the node or edge at the “source” of the Trans-omic interaction.

The 5th ~ 8th columns are about the target node.

The 1st and 5th columns are the network layer index.

The 2nd and 6th columns are whether the source and target of the interaction are node or edge.

The 3rd and 7th columns are the column name of the Cytoscape node or edge table.

The 4th and 8th columns are the string to be searched from the column with the name of the 3rd and 7th column. The Cytoscape node or edge that has the string will be the source or target of the Trans-omic interaction.

The last column is the type of the Trans-omic interaction. This information is added to the interaction column of the edge table.

3.2 Example

You can reproduce Figure5 of Yugi 2014 with the code below. This code execution will take some time to complete. (Do not operate Cytoscape until the code execution is completed.)

# suppressPackageStartupMessages(library(dplyr))
# suppressPackageStartupMessages(library(RCy3))
# suppressPackageStartupMessages(library(KEGGREST))
# Sys.setenv(LANGUAGE="en_US.UTF-8") 
library(transomics2cytoscape)
networkDataDir <- tempfile(); dir.create(networkDataDir)
networkLayers <- system.file("extdata", "yugi2014.tsv",
    package = "transomics2cytoscape")
stylexml <- system.file("extdata", "yugi2014styles.xml",
    package = "transomics2cytoscape")
suid <- create3Dnetwork(networkDataDir, networkLayers, stylexml)

layer1to2 <- system.file("extdata", "kinase2enzyme_gene2rea.tsv",
    package = "transomics2cytoscape")
suid <- createTransomicEdges(suid, layer1to2)

layer2to3 <- system.file("extdata", "allosteric.tsv",
    package = "transomics2cytoscape")
suid <- createTransomicEdges(suid, layer2to3)

Then, you should have a 3D view with layered networks and transomic interactions between them. (Note that you need to perform operations such as zooming out or adjusting the camera angle.)

Figure 3. Reproduction of Yugi’s Trans-omics visualizaion (Yugi 2014)

Figure 3. Reproduction of Yugi’s Trans-omics visualizaion (Yugi 2014)

4 Conversion from EC number to KEGG reaction ID

For those who have seen the enzyme reaction database such as BRENDA (Chang et al., 2021), it is not intuitive that the ID of the allosteric regulatory target (8th column of the allosteric.tsv) is the ID of the metabolic reaction rather than the EC number.

This is because KEGG uses the reaction ID instead of the EC number as the ID of the pathway object of the global metabolism map.

So transomics2cytoscape has a function ec2reaction that converts the EC number column of the Trans-omic interaction file into the KEGG reaction ID.

Figure 4. How the ec2reaction function works

Figure 4. How the ec2reaction function works

ecnum <- system.file("extdata", "allosteric_ecnumber.tsv",
    package = "transomics2cytoscape")
ec2reaction(ecnum, 8, "allosteric.tsv")

References

Chang, A., Jeske, L., Ulbrich, S., Hofmann, J., Koblitz, J., Schomburg, I., et al. (2021) BRENDA, the ELIXIR core data resource in 2021: New developments and updates. Nucleic Acids Research, 49, D498–D508.

Gehlenborg, N., O’Donoghue, S.I., Baliga, N.S., Goesmann, A., Hibbs, M.A., Kitano, H., et al. (2010) Visualization of omics data for systems biology. Nature Methods, 7, S56–S68.

Otasek, D., Morris, J.H., Bouças, J., Pico, A.R. and Demchak, B. (2019) Cytoscape Automation: Empowering workflow-based network analysis. Genome Biology, 20, 185.

Rohn, H., Junker, A., Hartmann, A., Grafahrend-Belau, E., Treutler, H., Klapperstück, M., et al. (2012) VANTED v2: A framework for systems biology applications. BMC systems biology, 6, 139.

Shannon, P. (2003) Cytoscape: A software environment for integrated models of biomolecular interaction networks. Genome Research, 13, 2498–2504.

Yugi, K., Kubota, H., Toyoshima, Y., Noguchi, R., Kawata, K., Komori, Y., et al. (2014) Reconstruction of insulin signal flow from phosphoproteome and metabolome data. Cell Reports, 8, 1171–1183.