TF_regulon_plot_from_scMEGA

#only run this if you are analysing mouse data 
grn$Target <- gsub("(?<=\\b)([a-z])", "\\U\\1", tolower(grn$Target), perl=TRUE)

#split the grn into tfs with regulons (target genes)
tfs <- unique(grn$Source)
grn_list <- list()
for (i in 1:length(tfs)){
grn_list[[i]] <- grn[grn$Source == tfs[[i]],]}

#chnage name (easier for plotting down stream)
name_tfs=paste("TF_", tfs, sep="")

#add module to seurat obj
for (i in 1:length(tfs)){
stia2021_rna <- AddModuleScore(stia2021_rna, features = list(grn_list[[i]]$Target), name = name_tfs[[i]] )}

#make heatmap
to_plot <- colnames(stia2021_rna@meta.data)[grep("TF_", colnames(stia2021_rna@meta.data))]

Idents(stia2021_rna) <- 'cluster.name' #chnage to the comparison you want todo
levels(stia2021_rna)[-c(2,3,12)]
stia2021_rna_f <- subset(stia2021_rna, idents=levels(stia2021_rna)[-c(2,3,12)])

dotplot<-DotPlot(stia2021_rna, features= to_plot, idents = levels(stia2021_rna)[-c(2,3,5)])

dotplot<-dotplot$data

dotplot<-dotplot %>% 
  select(-pct.exp, -avg.exp) %>%  
  pivot_wider(names_from = id, values_from = avg.exp.scaled) %>% 
  as.data.frame() 

dotplot$features.plot<-unique(dotplot$features.plot)
dotplot<-na.omit(dotplot)

row.names(dotplot) <- dotplot$features.plot  
dotplot <- dotplot[,-1] %>% as.matrix()

library(ComplexHeatmap)

Heatmap(dotplot, row_names_gp = gpar(fontsize = 5))