How to find whether a given network is scale free or not

1,128 views
Skip to first unread message

Kamalika Sen

unread,
Nov 5, 2015, 5:49:56 AM11/5/15
to cytoscape-helpdesk
Hi Scooter,

          As per your suggestion I ran clustermaker2 to identify the modules in my network and got the clusters. But, what I needed is the text file containing gene names of individual clusters and I didn't get it from there. I again ran allegro-mcode using Cytosacpe 2.8.0 and got the desired file.The visualization of individual clusters is also great. My query is that, is it ok if I use an older version (since allegro-mcode runs on that) instead of a newer one and if not than please tell me how can I get the desired text file in Clustermaker2.

Again, I need to know, how do I check whether a given network is scale free or not, is it modular and is it dense or not.

Regards
Kamalika

Alexander Pico

unread,
Nov 12, 2015, 1:25:08 PM11/12/15
to cytoscape-helpdesk
I'd recommend running Tools>Network Analyzer and assessing the results. You can also check out http://apps.cytoscape.org/apps/centiscape, which includes it's own tutorial.

Modularity and Density are not binary, i.e., you can say that they are or are not, but you can assess and compare the properties of two networks measured by the apps above. 

 - Alex

Scooter Morris

unread,
Nov 12, 2015, 1:32:12 PM11/12/15
to cytoscape...@googlegroups.com
Hi Kamalika,
    As Alex has suggested, using either Network Analyzer or Centiscape, then looking at the line fit through the degree distribution will tell you the extent to which the network is scale free.  None of the clustering algorithms will tell you that.  ClusterMaker assigns a cluster number to each cluster and writes that as an attribute for each node.  If you sort by that attribute, you can easily see all of the nodes that are part of each cluster.  You can get that into a text file by exporting the node attribute table and looking at it in Excel.  Also, I'm pretty sure Alex means to say that "you can NOT say that they are or are not,...".

-- scooter
--
You received this message because you are subscribed to the Google Groups "cytoscape-helpdesk" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cytoscape-helpd...@googlegroups.com.
To post to this group, send email to cytoscape...@googlegroups.com.
Visit this group at http://groups.google.com/group/cytoscape-helpdesk.
For more options, visit https://groups.google.com/d/optout.

Erno Lindfors

unread,
Jun 7, 2019, 9:10:31 AM6/7/19
to cytoscape-helpdesk
Hi Scooter, Alex and others,

I have been trying to find a feature for plotomg a degree distribution in Network Analyzer and Centiscape but so far without success. At this point perhaps it is good to take a look e.g. at https://en.wikipedia.org/wiki/Degree_distribution and https://mathinsight.org/degree_distribution to understand what the degree distribution means. Interestingly Network Analyzer has a "Node Degree Distribution" tab (e.g.  https://med.bioinf.mpi-inf.mpg.de/netanalyzer/index.php -> the first figure) but from what I understand this plot is not the probability distribution described in the above linked wiki pages; perhaps I would call it "Number of Nodes"-"Degree" scatter plot or so.

I have the same question as Kamalika has; I  want to investigate a given network is a scale-free. For this I think I need a probability distribution as stated at https://en.wikipedia.org/wiki/Power_law.

Please let me know if you know any way to do this in Cytoscape.
Also please feel free to correct me if you think I have misunderstood something.

I am using Cytoscape 3.7.1,  Network Analyzer 3.3.2 and CentiScaPe 2.2.

Best,
Erno
To unsubscribe from this group and stop receiving emails from it, send an email to cytoscape-helpdesk+unsub...@googlegroups.com.

Scooter Morris

unread,
Jul 11, 2019, 10:57:18 AM7/11/19
to cytoscape-helpdesk
Hi Erno,
   Actually, there is little difference between a node degree plot and a degree distribution plot from the perspective of calculating the power law.  The degree distribution is the same as the node degree plot, except that it is scaled by a constant value (the number of nodes).  Since the scale is constant across all of the points, it should not impact the power law calculation.

-- scooter
Reply all
Reply to author
Forward
0 new messages