Graphical presentation of STRING protein-protein network and gene ontology

Hello,

I found a paper (https://www.nature.com/articles/s41556-018-0220-2#Fig5) that they created STRING protein-protein network and also gene ontology bar charts as in their Figure 5.

Fig. 5: Mass spectrometry reveals the distinct reticular adhesome.

May I kindly ask you to let me have your valuable answers for my following questions please?

May I know how can I create the String network in such a way that it presents Direct binding of PtdIns(4,5)P2 in green nodes and Indirect PtdIns(4,5)P2 association in yellow nodes (as they did in their Figure 5)? Also, how they made some protein names with red text while the rest is black?

Also, I am highly interested to know how can I make gene ontology bar charts showing enriched terms along with -log10 (corrected P values) with red line, as they did.

I read STRING and DAVID databases but I could not find the options for making the above analyses and presentations. Any help to make these figures or similar ones with STRING, DAVID, or any other tools would be highly appreciated.

Many thanks. Best wishes,

string david gene-ontology interaction-network

• 7,055 views • link

updated 6.7 years ago by Leite

• written 6.8 years ago by Farah

Dear @F. Golestan

From my experience working with network (PPIN), I advise you to import data from cytoscape so you will have more flexibility to work by changing colors, font size, node color and whatever you want to change your layout.

With regards to DAVID, you can download GO result and build this graph using excel if you have no experience with R.

More information about cytoscape

https://manual.cytoscape.org/en/stable/

A: Cytoscape network of upregulate and downregulate transcription factors: how to d

A: How to multiple apply continuous mapping and Discrete mapping each column in Cyt

• 0 views • link

written 6.8 years ago by Leite

Dear Leite, Many thanks for your guide. Actually, before, I was using StringApp for pulling ppi networks. But, firstly, by using StingApp, I can not choose to retrieve only experimentally validated interactions (as it is possible in original String database). Secondly, I do not know how to detect "direct binding of protein X" and "indirect protein X association". Thirdly, I also do not know how to define sets and then map them to node fill colors (yellow and green).

Also, for creating this interaction network, do I need expression values or by having only protein names I can make such network with these features?

Also, is there any cytoscape app that I can obtain -log10 (corrected p values) for enriched GO biological process and KEGG pathways? Then, I can use them to make bar charts in Excel or R. I also tried ClueGO, but It does not calculate -log10 (corrected p values) which I need.

Sorry for asking several questions as I highly need help and guide. Thank you so much.

• 0 views • link

written 6.8 years ago by Farah

2 answers

Dear @F. Golestan,

Many questions, I will try to answer all.

First question:

I commonly use STRING db, and there I put my parameters as Interactions experimentally determined or from curated databases with a minimum interaction score of 0.400. Then I export txt and import it into cytoscape.

Second question:

If you had read supplemental material 3, you would have seen that he did it based on the literature.

Third question:

If you read these links you will understand how to change the color of nodes:

A: Cytoscape network of upregulate and downregulate transcription factors: how to d

A: How to multiple apply continuous mapping and Discrete mapping each column in Cyt

Also, for creating this interaction network, do I need expression values or by having only protein names I can make such network with these features?

You just need a list of genes or proteins.

DAVID and log10 (corrected p values)

The DAVID will give you the corrected P value (FDR), turn it into log10 scale and build your graph.

How can I convert -log10 (p-value) to p-value?

Excel LOG10 Function

• 0 views • link

written 6.8 years ago by Leite

Dear Leite, Thank you so much for your great help and guide. Your guides helped me a lot. As you suggested, I used STRING db with my parameters (only experiment, and confidence score of 0.4). I then exported TSV: tab separated values and imported into cytoscape (File -> Import -> Network from file...). However, in the created network in cytoscape, edges thickness is not shown by the confidence score of each edge (as it is shown in String db networks) which is the feature that I really need to make on my cytoscape network.

Also, I tried to change node shape to circle (Style -> Node -> Shape), but it seems that it does not have circle format for node shape, and it has Ellipse. But I need circle shape like String db.

Another feature that I really need to make on my cytoscape network, is to have node labels (protein names) exactly like String db, which each node name is written over a transparent box such that the edges under these transparent text boxes are also shown well (i.e. they are not hidden by the text box of protein names).

I will be very grateful if you can also give me your valuable help in the above issues. Many thanks.

• 0 views • link

written 6.8 years ago by Farah

Dear @F. Golestan

generate your network in string db using your list of genes or proteins as shown in figure 1.

Figure 1.

Make sure the string list is the same as the list you entered, then proceed Figure 2

Figure 2

3.Select your parameters Figure 3

Figure 3

Export your data Figure 4

Figure 4

Import to cytoscape Figure 5, figure 6 and figure 7

Figure 5

Figure 6

Figure 7

"Also, I tried to change node shape to circle (Style -> Node -> Shape), but it seems that it does not have circle format for node shape, and it has Ellipse. But I need circle shape like String db"

To do this select BIoPAX style, following the figure 8 and 9"

Figure 8

Figure 9

However, in the created network in cytoscape, edges thickness is not shown by the confidence score of each edge (as it is shown in String db networks) which is the feature that I really need to make on my cytoscape network.

To do this follow the steps in figure 10.

Figure 10

Now I think you have everything to work with, you just have to work hard and read the cytoscape manual and follow the tips I gave to you and in other questions.

Now it's up to you! Reading and using the software, everyone can learn with willpower and dedication.

Cytoscape 3.7.2 User Manual

A: Cytoscape network of upregulate and downregulate transcription factors: how to d

A: How to multiple apply continuous mapping and Discrete mapping each column in Cyt

Best regards Leite

• 0 views • link

written 6.7 years ago by Leite

Dear Leite, Thank you so much for your great help and support. That was perfect. Your help and patience is highly appreciated. Best regards.

• 0 views • link

written 6.7 years ago by Farah

Dear Leite, Thank you very much for your great help and explanation. I could do it perfectly. But, I only could not find how to make node style like 3D nodes by adding both black shadow and white specular highlight over nodes (exactly like String database style).

So, after importing the interaction output of String database which is in TSV format into Cytoscape, I used BioPAX style for my imported network. However, I really need to make my nodes like 3D nodes similar to Stringdb node style. I searched and read Cytoscape tutorials, but I could find its solution. I would highly appreciate if you help and advise me that how to add both black shadow and white specular highlight over nodes in such a way that I can also modify, everything within this node style such as node color, size,...)? Thank you so much. Best wishes

• 0 views • link

written 6.7 years ago by Farah