Help with basics for graph visualization

[bumper_boy2000]

I've scanned through the forums, and haven't found any definite guide to achieve what I'm expecting.

I have attached a file below. The file name is corum.csv which contains data from CORUM.

How the data looks like is as follows:
C-CFTR2-NHERF-ezrin#complex;P13569;P15311;O14745
E.F.G#complex;P62304;P62306;P62308
12S#U11#snRNP;Q9BV90;Q6IEG0;Q8N8D1;Q9H875;Q16629;P14678;P62314;P62316;P62318;P62304;P62306;P62308;Q96GM8;Q16560;P67809
13S#condensin#complex;Q15021;Q9BPX3;Q15003;O95347;Q9NTJ3
17S#U2#snRNP;Q8IWX8;Q7L014;O43143;O75937;Q9P0W2;P10809;Q7RTV0;Q9UHX1;Q96I25;Q15459;Q15428;Q12874;O75533;Q9Y3B4;Q13435;Q15393;Q15427;Q9BWJ5;Q07955;O75940;P09661;P14678;P08579;P62314;P62316;P62318;P62304;P62306;P62308;O15042;Q12931;Q01081;P26368
18S#U11/U12#snRNP;Q9BV90;Q6IEG0;O43143;Q8N8D1;Q6P2Q9;Q96LT9;O75533;Q9Y3B4;Q13435;Q15393;Q15427;Q9BWJ5;P14678;P62314;P62316;P62318;P62304;P62306;P62308;Q16560;P67809;Q8TBF4;Q9UDW3;Q15696
20S#methylosome#and#RG-containing#Sm#protein#complex;P54105;O14744;P14678;P62314;P62316;P62318
20S#methylosome-SmD#complex;P54105;O14744;P62314;P62318
20S#methyltransferase#complex;P54105;O14744;Q9BQA1
20S#methyltransferase#core#complex;P54105;O14744
...
...
...

These are just the first 10 lines.
Each line is semicolon-delimited. (the # symbols are part of the complex name)
The first element is the COMPLEX, while the subsequent elements are the SUBUNITS belonging to the complex.
What I'm looking for as an end result is for the graph to visualize so that it's possible to differentiate between a COMPLEX and a SUBUNIT when viewed. Now, everything looks like they're equally unknown in the graph. There's no distinct indication to see that for example this node is a COMPLEX or a SUBUNIT. Attached here also is a screenshot of my current visualization. I have no idea how to set the settings in order to achieve that.

My minimal settings are as circled in the screenshot-corum.png

A zoomed in view is provided at screenshot-corum-zoom.png (you should see where I'm coming from)

Hope to receive some input and help! Thanks!

[phreakocious]

If you convert your input data from that format to Gephi's GDF format, you can have a 'nodedef' section at the top that specifies a different size or color for those nodes that you want to highlight.

It looks something like this:

Code: Select all

nodedef> name, fname VARCHAR(128), label VARCHAR(80), relation VARCHAR(16),conns INT, width DOUBLE, color VARCHAR,labelvisible BOOLEAN
v1,"phreakocious","/phreakocious",xXx,2820,65,'255,215,0',true
v2,"2Tuff","/2tuff",other,1987,20,'249,232,176',true
v3,"8-Bit","/8-bit",other,201,20,'249,232,176',true

edgedef> node1,node2,directed BOOLEAN,weight DOUBLE
v1,v2,true,1
v1,v3,true,1
v2,v3,true,1

Does that help at all?

[phreakocious]

To add to the previous thought, you could also add another variable where you designate each node as 'complex' or 'subunit' . This would allow you to use the filters within Gephi to select them and adjust these parameters later.