Task 3 - Dynamic Network Visualization

DiNAR is a tool used to examine omics datasets in a prior knowledge context. This is achieved through the dynamic representation of biological data over multiple timepoints or conditions in a network format.

The current task will guide the user on how to import and analyse networks into DiNAR and obtain outputs like these.

To start using DiNAR, two options are provided in this usecase to prepare a gene/protein network with specific requirements:

* Option 1 - Using the provided Cytoscape_to_Dinar.R script to prepare the network.

* Option 2 - Using the DiNAR preprocessing subApp to prepare the network.

Option 1 - Importing a Custom Network into DiNAR with Cytoscape_to_Dinar

This option allows the user to provide simple node and edge tables to import into DiNAR. For example, these can be obtained from Cytoscape (access this section to learn how).

Additional settings are provided to enrich the network with the corkoak.results obtained in Task 1 - Plant Data Annotation, and usefull information from the Ensembl Plants database.

For this option, the initial node and edge tables should have the following format(s):

• Node Table Example (.csv):

geneID
CFP56_57155
CFP56_18234
CFP56_55251
…

Note

This column is mandatory in the first position and must have gene/protein IDs. Additional columns can be present.

• Edge Table Example 1 (.csv):

geneID1	geneID2	reactionType
CFP56_57155	CFP56_18234	coexpressed_with
CFP56_57155	CFP56_55251	coexpressed_with
…	…	…

• Edge Table Example 2 (.csv):

name	interaction
CFP56_57155 (interacts with) CFP56_18234	interacts with
CFP56_57155 (interacts with) CFP56_55251	interacts with
…	…

Note

The corkoak_node.csv and corkoak_edge.csv, present in the CorkOak_UseCase_Data-main folder, already conform to this format.

1. Open the command line (Command Prompt) and run Cytoscape_to_Dinar.R

   For a detailed usage of this script, go to Cytoscape_to_Dinar Usage.

   #Don't forget to replace the PathToFolder section with the Directory where the CorkOak_UseCase_Data-main Folder is located
   Rscript PathToFolder/Cytoscape_to_Dinar.R -n corkoak_node.csv -e corkoak_edge.csv -c corkoak.results.txt -o corkoak
   

After running the script, the output files should look like the following:

• DiNAR Ready Node Table corkoak_dinar_node.txt:

geneID	shortName	shortDescription	MapManBIN	clusterID	x	y	clusterSimplifiedNodeDegree	expressed
CFP56_57155	POF01545.1	-	21.4.1.1.2.2_Cell wall organisation.cell wall proteins.hydroxyproline-rich glycoprotein activities.arabinogalactan-protein activities.Fasciclin-type arabinogalactan protein activities.Fasciclin-type arabinogalactan protein *(FLA11)	1	20.0223811911811	7.2089051057372	12	1
CFP56_18234	POE85887.1	-	21.2.2.1.3.2_Cell wall organisation.hemicellulose.xylan.biosynthesis.xylosyltransferase activities.xylosyltransferase *(IRX10)	1	24.1097974399641	9.36531030251287	10	1
CFP56_55251	POE60447.1	-	21.1.2.2_Cell wall organisation.cellulose.cellulose-hemicellulose network assembly.regulatory protein *(COB)	1	18.0888090500896	6.01548035406399	12	1
…	…	…	…	…	…	…	…	…

• DiNAR Ready Edge Table corkoak_dinar_edge.txt:

geneID1	geneID2	reactionType	clusterID_geneID1	clusterSimplifiedNodeDegree_geneID1	clusterID_geneID2	clusterSimplifiedNodeDegree_geneID2	exists
CFP56_57155	CFP56_18234	coexpressed_with	1	12	1	10	1
CFP56_57155	CFP56_55251	coexpressed_with	1	12	1	12	1
…	…	…	…	…	…	…	…

With the tables ready, jump to the following Main DiNAR section.

Option 2 - Preparing a Custom Network into DiNaR (from other sources)

Alternatively, to import a custom network into DiNAR and avoid the Cytoscape_to_Dinar.R script, a node and edge tables should be supplied following a specified tabular structure, which must be firstly processed with the DiNaR preprocessing subApp. The table formats necessary for this option are the following:

• Node Table Example (before DiNAR preprocessing):

geneID	shortName	shortDescription	MapManBin
CFP56_57155	-	-	-
CFP56_18234	-	-	-
CFP56_55251	-	-	-
…	…	…	…

The first column is the only one required to be filled (either with gene or protein IDs). The other columns are required but may be left empty (using “-“).

Additional columns can also be present.

• Edge Table Example (before DiNAR preprocessing):

geneID1	geneID2	reactionType
CFP56_57155	CFP56_18234	coexpressed_with
CFP56_57155	CFP56_55251	coexpressed_with
…	…	…

1. Open the DiNaR preprocessing subApp and select the tables option tab

2. Click Choose Nodes File and select a input node table

3. Click Choose Edges File and select a input edge table

4. Press Go!

5. In the Nodes Tab, click Download (button under the node table). It should look like the following:

geneID	shortName	shortDescription	MapManBIN	clusterID	x	y	clusterSimplifiedNodeDegree	expressed
CFP56_57155	-	-	-	1	18.8867566488289	7.80493942833234	12	1
CFP56_18234	-	-	-	1	21.8082235670472	6.47746918475539	10	1
CFP56_55251	-	-	-	1	22.2041564614846	13.1541499502343	12	1
…	…	…	…	…	…	…	…	…

6. In the Edges Tab, click Download (button under the edge table). It should look like the following:

geneID1	geneID2	reactionType	clusterID_geneID1	clusterID_geneID2	clusterSimplifiedNodeDegree_geneID1	clusterSimplifiedNodeDegree_geneID2	exists
CFP56_57155	CFP56_18234	coexpressed_with	1	1	12	10	1
CFP56_57155	CFP56_55251	coexpressed_with	1	1	12	12	1
…	…	…	…	…	…	…	…

With the tables ready with this alternative method, we can then import them into DiNAR.

Video guide for DiNaR preprocessing subApp

Importing a Custom Network into the Main DiNaR App

1. Go to the DiNaR App

2. In select network, choose Custom network

3. In Upload nodes table, select the corkoak_dinar_node.txt file

4. In Upload edges table, select the corkoak_dinar_edge.txt file

5. In Upload Experimental data files, select the corkoak_LogFC_April.txt, corkoak_LogFC_June.txt and corkoak_LogFC_July.txt files

6. Click the Files selection and order box → press backspace twice → select corkoak_LogFC_June and then corkoak_LogFC_July

7. Click Proceed!

Exploring DiNaR network visualization options

1. Click the Cluster Information Tab, and select the MapMan BINs option

This image highlights the most populated Bins throughout the Input Network, with the BIN-35.2_not assigned.not annotated being the most prevalent due to the abundance of uncharacterized genes in this species.

The BIN-21.6.2.2_Cell wall organisation.lignin.monolignol conjugation and polymerization.lignin laccase, a important process of lignin biosynthesis, hints to the specific role of the example gene set acting on tree secondary growth and development.

Creating a dynamic network visualization

Note

Larger amounts of experimental data allows the user to create longer and richer dynamic network visualizations in DiNAR.

1. Click the Differential expression per cluster Tab, and select the Dynamic-animatoR option

2. Scroll down and, under the Select speed option, click on the video speed 1

Note

Different speed settings can be used to obtained visualizations which go faster (speed 1) or slower (speed 0.1) throughout multiple conditions/timepoints.

3. Click Create dynamic.html

4. After a couple seconds (depending on network size), click Download dynamic.html

5. Open the .html file and click the starting video symbol

• Cork Oak Dynamic Network Visualization:

It is clear to notice, through the visualization of the generated video, that all cork oak genes present in the example network have a concerted increase in gene expression towards the month of July (summer). Together with the information inferred in the previous task, these observations highlight a steady gene upregulation in cork oak as this species seasonal growth includes gene upregulation towards lignin biosynthesis and tree growth processes during summer.

Larger and more complex networks, with mixing gene downregulation and upregulation, and additional curated data, can be visualized in DiNAR to observe time-specific phenomena in a clear and intuitive manner.

Video guide for DiNaR App

Now, if ready, the user can move into the next and final Task 4 - Ensembl Plants Query, which will help in the retrieval of information from the Ensembl Plants database from the command line.

Getting Node and Edge Tables from Cytoscape

Cytoscape is a widely used tool for network analysis and visualization. This section will guide the user on how to obtain node/edge tables to be imported into DiNAR using the Cytoscape_to_Dinar.R script.

1. In Cytoscape, select the desired network

2. Click the Node Table tab

3. Select the Export Table to File… option, saving it in the default (.csv) format

4. Click the Edge Table tab

5. Select the Export Table to File… option, saving it in the default (.csv) format

Note

Both node and edge tables can be retrieved from several databases or retrieved from other network visualization tools, as long the structure remains DiNAR compatible.

With these tables from Cytoscape, the user can resume the workflow to prepare and import them into DiNAR here.

Rscript Usage:

Rscript PathToFile/Cytoscape_to_Dinar.R [-n] node_table.csv [-e] edge_table -c corkoak.results -e species_name -o output_name

• [-n] node table - Input a Node Table (.csv, comma-separated)

• [-e] edge table - Input a Edge Table (.csv, comma-separated)

• [-c] custom annotation - Input gene/protein annotations obtained in Mercator4 / mercator.results (optional)

• [-e] ensemblplants - Retrieve short names, short descriptions and orthologs from the Ensembl Plants database (optional)

• [-o] output - Specify a name for the output node and edge tables

Warning

Do not forget to specify the path for the directory where this script is located (CorkOak_UseCase_Data-main). This should also be the folder with the corkoak_node.csv, corkoak_edge.csv and corkoak.results.txt files.

DiNAR Output Examples

The following visualizations were obtained from a Arabidopsis thaliana network showcasing gene expression in multiple timepoints after Pseudomonas syringae infection (described here).

• Example 1 - Interactive Visualization Option (.html format)

This option allows the user to click nodes and edges of the network to check their summarized information, and highlights the expression tendencies of specific genes throughtout the course of infection response.

• Example 2 - Video Option (.mp4 format)

This concludes Task 3 of the current usecase.

The next and final Task 4 - Ensembl Plants Query will guide the user on how to retrieve information from the Ensembl Plants database, independently of the Cytoscape_to_Dinar.R script.