search

jinworks / CellChat

R toolkit for inference, visualization and analysis of cell-cell communication from single-cell and spatially resolved transcriptomics
GNU General Public License v3.0
195 stars 21 forks source link

What Communication probability value is to be considered for a believable interaction? #152

Open tumorscholar opened 1 month ago

tumorscholar commented 1 month ago

Hello What Communication probability value is to be considered for a believable interaction?
Vignette https://htmlpreview.github.io/?https://github.com/jinworks/CellChat/blob/master/tutorial/CellChat-vignette.html Figure 'CXCL signaling' heatmap legend has a max of communication probability 0.15, meaning 15% of finding this interaction true, if I understood correctly.

luxhexmaster commented 3 weeks ago

Hi Prof. Jin,

I understand that based on RNA expression values, CellChat computes a Communication Probability and a p-value for each R-L interaction. To my understanding, the p-value indicates whether the interaction is believable, but I'm not sure what to make of the Communication Probability value. I had two questions on this:

1.) As @tumorscholar asks, assuming all p-values meet the 0.05 threshold, what Communication Probability value is high enough for the interaction to be believable?

2.) When comparing Communication Probability values across conditions, what difference in Communication Probability would be high enough for us to hypothesize that this R-L interaction is different between groups? For example, if CDH1 --> CDH1 interactions are 0.06 and 0.12 between conditions A and B, would that be considered significantly different?

Any guidance on this would be greatly appreciated. Thank you for developing this package!

sqjin commented 3 weeks ago

@luxhexmaster @tumorscholar We did not set any threshold for the communication probability, which is often a common way in the field of cell-cell communication analysis. But we can rank the interaction based on the interaction probability once they are significant enough based on the p-value. The " communication probability 0.15" means the interaction strength is 0.15.

@luxhexmaster When you compare two conditions, we also provide the identification of dysfunctional signaling by combing with the differential expression analysis, where we provide a p-value.

luxhexmaster commented 3 weeks ago

Thank you very much Prof. Jin