BaderLab / EnrichmentMapApp

The EnrichmentMap Cytoscape App allows you to visualize the results of gene-set enrichment as a network.
http://apps.cytoscape.org/apps/enrichmentmap
GNU Lesser General Public License v2.1
31 stars 12 forks source link

gray nodes (not colored by -log10(value)when creating the map #540

Closed veroniquevoisin closed 4 months ago

veroniquevoisin commented 4 months ago

gray nodes (not colored by -log10(value)when creating the map I then modify the fill color style myself. AA was working fine.

Screenshot 2024-02-29 at 8 02 35 AM Screenshot 2024-02-29 at 8 03 02 AM
mikekucera commented 4 months ago

Hi Veronique, can you please send me the session file. Thanks.

veroniquevoisin commented 4 months ago

Hi, I got a new one today. It is again when I create a new map on a complex session with already a lot of maps (not sure if related).

Screenshot 2024-03-06 at 4 51 39 PM
veroniquevoisin commented 4 months ago

link to session file: https://www.dropbox.com/s/8g5z6v0emckyi9q/test%20_forMike.cys?dl=0

mikekucera commented 4 months ago

Hi Veronique,

In the network above the entire p-value column is zeros. The -log10(pvalue) column is blank. That causes the mapping to not have a range which results in the grey nodes. Is there something wrong with the data, or is this a valid case?

Regardless it should detect when the range is invalid for -log10(pvalue) and use a different mapping.

veroniquevoisin commented 4 months ago

Hi Mike, Yes, it is a valid case. I have an example right now where all the pvalue are 0. They are so many significant pathways that I do a map with the top pathways and GSEA returned them with a pvalue = 0 and -log10(0) = inf. In this case, all nodes should be red and I suppose that you can replace the 0 by any small value? eg 1e-10?

Screenshot 2024-03-07 at 10 28 20 AM

mikekucera commented 4 months ago

Hi Veronique.

When all the pvalues are 0.0 (or missing) then it uses 1e-10 for the pvalue. This results in 10.0 for -log10(pvalue).

This is an example of what the network and style look like when all pvalues are 0.0. Is this ok?

Screenshot 2024-03-07 at 11 22 26 AM
veroniquevoisin commented 4 months ago

Yes, it looks ok