search

saezlab / decoupleR

R package to infer biological activities from omics data using a collection of methods.
https://saezlab.github.io/decoupleR/
GNU General Public License v3.0
176 stars 23 forks source link

Question about TF activity inference analysis in single cell integrated datasets #111

Closed ChrisTzaferis closed 5 months ago

ChrisTzaferis commented 5 months ago

Hi, thank you for developing this very useful package! I would like to ask you for some clarifications regarding the TF activity inference on single cell datasets.

1) In case the single cell Seurat object contains more than one conditions after integration or batch correction analysis, is it correct to follow the vignette using the data slot of the RNA assay or it would be advisable to perform the analysis separately for each condition?

2) If it is correct to proceed with the analysis in the integrated object as described above, can the methodology described for the heatmap production (detecting the top deviant TFs) be utilized to compare activities between the clusters of the dataset, when they follow the format: cluster1_Healthy, cluster1_Disease, cluster2_Healthy, cluster2_Disease,...,etc?

Thank you for your time.

PauBadiaM commented 5 months ago

Hi @ChrisTzaferis,

If your aim is to compare conditions, I would rather perform differential expression analysis at the pseudobulk level, per cell type and then take the contrast statistics that you obtain (for example the stat column of DESeq2) as input for decoupleR. This way you would obtain which TFs are differentially changing the most for each cell type. Once you have the contrasts, you can follow the vignette for bulk, else we have a pseudobulk vignette but it is in python. Hope this is helpful!

ChrisTzaferis commented 5 months ago

Thank you for your suggestion, I will give it a try!