Question: Cytoscape-gene expression data
gravatar for Malgorzata.Wieteska
2.1 years ago by
Malgorzata.Wieteska10 wrote:

Dear All,

I have time series expression data. I have visualised their values on the network. However, I'm not sure what is the best approach to interpret the results in the context of a biological network. I wonder if I should visualise the expression value at the given time or log2 transform data and check the FC between two different experimental conditions at a given time or the difference in FC between the same experimental conditions at two different times. I try to follow the basic expression analysis tutorial but this issue is not described in details and I don't know what is the best approach in order to infer if a transcriptional activation activity of the specific gene is repressed by some other genes.

In the tutorial they have written that "The gal80Rexp expression values will be mapped to node color; nodes with low expression will be colored blue, nodes with high expression will be colored red." but later they had said :"Both nodes (GAL4 and GAL11) show fairly small changes in expression, and neither change is statistically significant: they are pale blue with thin borders. " Thus, I'm puzzled.

I will appreciate any suggestion.

ADD COMMENTlink modified 17 months ago by alfonsoolayaabril0 • written 2.1 years ago by Malgorzata.Wieteska10

Good afternoon, I am trying to do a proteomic analysis of differential expression in Cytoscape. I'm looking at the tutorial and I've made the network of my proteins using STRING, but when I have to enter the differential expression data following the instructions in the tutorial, I create the column, but without data. The data that I intend to include are those of fold change, and although I have them initially in Excel, I save them as .csv. Do you know what may be happening that does not load the data of fold change? I have tried the fold change data as such, in log2, log10, removing spaces everywhere, etc ...

ADD REPLYlink written 17 months ago by alfonsoolayaabril0
gravatar for xanderpico
2.1 years ago by
xanderpico300 wrote:

Hi. There are a few directions you could go with your analysis of expression data in Cytoscape. The first step in most cases is to identify the source of the network or pathway that you want to work with. Unfortunately, there is not a single "true" network :). Souces include:

  • NDEx - for user network collected from databases and users
  • STRING - for networks constructed from a query gene set (e.g., the significantly differentially expressed genes from your data)
  • GeneMANIA - for networks constructed from a query gene set
  • WikiPathways - for collected pathway models with mechanistic detail queriable by name or genes
  • clusterMaker2 - for a coexpression network constructed from your own data (see Create Correlation Network)
  • many other sources...

Then the tutorial you found kicks in to descibe how to merge your data with a network in Cytoscape and control viualization. The interpretation of the data in the context of networks, however, is not done by an app... for that humans are still needed :). Folks with domain expertise for specific types of biology still have an important job.

There are also collections of apps that suggest various analysis routes:

More tutorials are being added all the time at

ADD COMMENTlink written 2.1 years ago by xanderpico300
gravatar for erdiazval
20 months ago by
erdiazval50 wrote:

Hi Malgorzata I think you may have a misunderstanding in terms of what a co-expression network is and how they are built. Let's say you have 2 experimental conditions across 6 developmental stages with 3 replicates each; this would give you a 36 samples experiment. If you want to make a network-based analysis then, you would ask the data what genes display correlated expression profiles. The biological meaning would be for example: if a transcriptional repressor is up-regulated in a experimental condition, then their targets downstream would be down-regulated. One simple solution to achieve this task that I've been able to perform would be: 1) Normalize expression profiles via DESeq2/edgeR for the 36 samples 2) Generate co-expresssion matrix via AracNe algorithm (I could share my scripts) 3) Import co-expression matrix into cytoscape and build network(s) 4) Import log2 FoldChange of experimental conditions for each gene in your matrix 5) Map log2FC into nodes in a heatmap fashion

This approach will allow you to know how expression variation affecting important genes (central hubs) diffuses across the genetic network.

ADD COMMENTlink written 20 months ago by erdiazval50
Please log in to add an answer.


Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 2174 users visited in the last hour