scRNA-seq_analysis

pipelines/23_super_markers/super_markers.R

@@ -0,0 +1,162 @@
+args = commandArgs(trailingOnly=T)
+seurat.addr = args[1]
+set.ident = "cell.labels"
+
+passed.threshold         = .5
+competition.threshold    = .8
+expansion                = 4
+    
+library(Seurat)
+library(plyr)
+library(dplyr)
+
+source("../../tools/bunddle_utils.R")
+
+seurat.addr = file.path("../../data", seurat.addr)
+
+gene_mean_expression = function(seurat.obj){
+  no.genes = nrow(seurat.obj@data)
+  start_index = 1
+  while (start_index < no.genes){
+    end_index = start_index + 999
+    end_index = min(end_index, no.genes)
+    expression.data_ = data.matrix(seurat.obj@data[start_index:end_index, ])
+    expression.data_ = t(expression.data_)
+    expression.data_ = as.data.frame(expression.data_)
+    expression.data_ = cbind(data.frame(CellLabels = as.vector(seurat.obj@ident)), expression.data_)
+    expression.data_ = aggregate(expression.data_[2:dim(expression.data_)[2]], list(expression.data_$CellLabels), mean)
+    expression.data_ = cbind(data.frame(CellType = expression.data_$Group.1), expression.data_[, 2:dim(expression.data_)[2]])
+    rownames(expression.data_) = expression.data_$CellType
+    expression.data_ = expression.data_[, 2:ncol(expression.data_)]
+    print(start_index)
+    if (start_index == 1){
+      expression.data = expression.data_
+    }else{
+      expression.data = cbind(expression.data, expression.data_)
+    }
+    start_index = start_index + 1000
+  }
+  expression.data
+}
+
+print("Loading data ... ")
+seurat.obj = readRDS(seurat.addr)
+seurat.obj = SetAllIdent(seurat.obj, id=set.ident)
+print("Data loaded.")
+
+# get aggregated expression matrix
+expression.data = gene_mean_expression(seurat.obj)
+saveRDS(expression.data, "gene_mean_expression.RDS")
+
+#
+gene.names = colnames(expression.data)
+cell.types = rownames(expression.data)
+n.cell.types = length(cell.types)
+super.markers = rep("", length(cell.types))
+names(super.markers) = cell.types
+
+gene.max.expression  = apply(expression.data, 2, max)
+expression.data.norm = t(t(expression.data) / gene.max.expression)
+
+for(j in seq_along(gene.names)){
+  gene.name = gene.names[j]
+  super.vector = expression.data.norm[, gene.name]
+  super.order  = order(super.vector, decreasing=T)
+  cell.type    = names(super.vector[super.order[1]])
+  expr.val     = as.vector(expression.data[cell.type, gene.name] )
+  competition = all(super.vector[super.order[expansion:length(super.order)]] < competition.threshold)
+  if(competition & expr.val > passed.threshold){
+    super.markers[cell.type] = paste(super.markers[cell.type], gene.name, sep = ", ")
+  }
+}
+
+# order markers
+for (i in seq_along(super.markers)){
+  ms        = super.markers[i]
+  cell.type = names(ms)[1]
+  ms = unlist(strsplit(as.character(ms), split=", "))
+  ms = ms[ms != ""]
+  expression.row = expression.data[cell.type, ms]
+  ms = ms[rev(order(expression.row))]
+  ms = paste(ms, collapse = ", ")
+  super.markers[i] = ms
+}
+
+# make classification markers
+classification.markers = super.markers
+
+seurat.obj = FindVariableGenes(object = seurat.obj, mean.function = ExpMean, 
+                               dispersion.function = LogVMR, x.low.cutoff = .0125, 
+                               x.high.cutoff = 9, y.cutoff = .2)
+
+for (k in seq_along(cell.types)){
+  cell.type = cell.types[k]
+  print("printing cell.type")
+  print(cell.type)
+  eval(parse(text = sprintf("cells.markers = super.markers[['%s']]", cell.type)))
+  cells.markers = unlist(strsplit(cells.markers, split=", "))
+  print("printing cell.markers")
+  print(cells.markers)
+  cor.expression.data = expression.data[, cells.markers]
+  print("printing cor.expression.data")
+  print(cor.expression.data)
+  compare.to = cell.types[cell.types != cell.type]
+  print("printing compare.to")
+  print(compare.to)
+  t.cor.expression.data = t(cor.expression.data)
+  colnames(t.cor.expression.data) <- cell.types
+  rownames(t.cor.expression.data) <- cells.markers
+  print("printing t.cor.expression.data")
+  print(t.cor.expression.data)
+  cor.dist = cor(t.cor.expression.data, method="spearman")
+  rownames(cor.dist) <- cell.types
+  colnames(cor.dist) <- cell.types
+  print("printing cor.dist")
+  print(cor.dist)
+  cor.dist = cor.dist[cell.type, compare.to]
+  print("printing cor.dist[cell.type, compare.to]")
+  print(cor.dist)
+  compare.to = names(cor.dist)[cor.dist > .3]
+  print("printing length of compare.to")
+  print(length(compare.to))
+  if(length(compare.to) > .45){
+    cor.dist = cor.dist[compare.to]
+    cor.dist = cor.dist[order(cor.dist, decreasing=T)]
+    compare.to = names(cor.dist)[1:3]
+  }
+  if (length(compare.to) > 0 & !any(is.na(compare.to))){
+    print(sprintf("Comparing %s to: %s", cell.type, paste(compare.to, collapse = ", ")))
+    DEGs = list()
+    for (m in seq_along(compare.to)){
+      DEG = FindMarkers(object=seurat.obj, ident.2=compare.to[m], ident.1=cell.type,max.cells.per.ident=500, only.pos=T, min.pct=.5, genes.use=seurat.obj@var.genes, )
+      DEG$cluster = compare.to[m]
+      DEG$gene = rownames(DEG)
+      DEGs[[m]] = DEG
+    }
+    DEGs = Reduce(f=rbind, x=DEGs)
+    additional.markers = (DEGs %>% group_by(cluster) %>% top_n(5, avg_logFC))$gene
+    new.markers = unique(c(cells.markers, additional.markers))
+    new.markers = paste(new.markers, collapse = ", ")
+    eval(parse(text = sprintf("classification.markers[['%s']] = new.markers", cell.type)))
+  }
+}
+
+
+# order markers
+for (i in seq_along(classification.markers)){
+  ms        = classification.markers[i]
+  cell.type = names(ms)[1]
+  ms = unlist(strsplit(as.character(ms), split=", "))
+  ms = ms[ms != ""]
+  expression.row = expression.data[cell.type, ms]
+  ms = ms[rev(order(expression.row))]
+  ms = paste(ms, collapse = ", ")
+  classification.markers[i] = ms
+}
+
+saveRDS(classification.markers, "signatures.RDS")
+
+sig.df = data.frame(CellTypes = names(classification.markers), Signatures = classification.markers)
+write.csv(sig.df, "Signatures.csv", row.names = F)
+
+print("Ended beautifully ... ")

pipelines/23_super_markers/super_markers.sh

@@ -0,0 +1,11 @@
+#!/bin/bash
+
+#$ -cwd
+#$ -N super_markers
+#$ -V
+#$ -l h_rt=23:59:59
+#$ -l h_vmem=400G
+
+Rscript super_markers.R $1
+
+echo "End on `date`"