I want to analyze how differential expressed (DE) genes in Reactome pathways (R-HSAs) are connected to active transcription factors (TF) to identify potential co-regulation and better understand the role of TFs in the R-HSAs. I believe this is called gene regulatory network (GRN) analysis. With the results I want to construct and visualize a network diagram showing the genes/TFs as nodes clustered according to the R-HSAs and their connections as edges. However, I have to continue working with the results of an already completed DE gene analysis because the computation was done in a complicated way, using sampling weights.
What I have:
- human TMM-normalized mRNA-seq data from blood
- DE gene analysis,
log2FCas well as
p.adjavailable in a result matrix for each ENSEMBL gene (~17K, ~1K DE)
- from DE genes I obtained some up- and downregulated functional R-HSAs and 27 active TFs (and TF:TF interactions) using
I manually created overlaps between the active TFs and the DE genes of the individual R-HSAs, which was insightful, but of course it doesn't give me any stats.
Following biostars answers, I have tried a number of packages, such as
CeTF, but 1. they do not correspond to the intention of the analysis and 2. they can't include pre-analyzed DE gene data with sampling weights.
Then I tried
RTN, which looks promising, but I'm not sure how to use the R-HSAs in it. I also tried
SBGNviewon the R-HSAs, which looks beautiful, but I'm not sure how to 1. combine more R-HSAs in one plot with co-expressed genes connected, or 2. show TFs there.
Seeking advice on:
- tools or methods that can integrate my existing DE genes, R-HSAs, and TF data for detailed GRN analysis
- strategies for creating/visualizing a network that highlights TFs co-regulating the R-HSAs
- good example studies that have done something similar