Master regulators in LUAD

Workflow Surgical Resection EGFR TKI Immunotherapy Data Citation

A Robust and Specific Workflow to Identify Master Regulators

---

Protein activity

ZWINT target genes

source("code/MR.R")
load("Data/LungMR_GRN.RData")

ExS = GEX[["TCGA"]]
dM = ExS - apply(ExS, 1, mean)

regulon = GRN[["TCGA"]]
tfmode = regulon$ZWINT$tfmode
gs = list(up = sort(names(tfmode)[tfmode > 0]), dn = sort(names(tfmode)[tfmode < 0]))
gs

$up
 [1] "CCNB2"     "CDC25C"    "CDC7"      "CDCA8"     "CDK1"      "EZH2"     
 [7] "KIF11"     "KIF23"     "LINC00471" "NCAPG"     "NEK2"      "NUSAP1"   
[13] "POLE2"     "PRC1"      "PSRC1"     "RAD51"     "SGO1"      "SKA3"     
[19] "SPDL1"     "TAF5"      "TIGD2"     "TXNDC5"    "VRK1"      "WRN"      

$dn
[1] "ARRDC4"  "C9orf47" "FAM117A" "MAOB"    "RIMKLA"  "ZDHHC9" 

ZWINT activity

sampleName = c("TCGA-50-5072", "TCGA-44-6148")

for(i in 1:2){
  signature = dM[, sampleName[i]]
  p = plotEnrichment(gs[["up"]], signature) + theme_bw() + labs(title= sampleName[i]) + 
    theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank())
  print(p)
}

Robustness

Signatures

T1 = read.table(file = "Signature/Xie.txt", row.names = 1, header = TRUE)
xie = list(HihgRisk = filter(T1, HR_FFPE > 1)$Symbol, LowRisk = filter(T1, HR_FFPE < 1)$Symbol)

ES = as.character(read.table(file = "Signature/ES_exp_1.txt", 
  row.names = NULL, header = TRUE, sep = "\t")[,1])
PRC2 = as.character(read.table(file = "Signature/PRC2_targets.txt", 
  row.names = NULL, header = TRUE, sep = "\t")[,1])
gs = list(ES = ES, PRC2 = PRC2)

cat("Number of genes shared between ESC and Xie:", length(intersect(xie, ES)), "\n")

Number of genes shared between ESC and Xie: 0 

Master regulator identification

m1 = mrMulti(GEX, GRN, xie, nsample = 50)
write.table(m1, file = "out/LungMR_Xie.txt", row.names = FALSE, col.names = TRUE, sep = "\t")

m2 = mrMulti(GEX, GRN, gs, nsample = 50)
write.table(m2, file = "out/LungMR_ESC.txt", row.names = FALSE, col.names = TRUE, sep = "\t")

Comparison

m1 = read.table(file = "out/LungMR_Xie.txt", row.names = NULL, header = TRUE, sep = "\t")
m1$signature = "Xie"
m2 = read.table(file = "out/LungMR_ESC.txt", row.names = NULL, header = TRUE, sep = "\t")
m2$signature = "ESC"
mergedT = rbind(m1, m2)

uTable = filter(mergedT, p.value < 0.01, NES > 0)
dTable = filter(mergedT, p.value < 0.01, NES < 0)

x = table(uTable$Regulon, uTable$signature)
y = as.data.frame.matrix(x)
z = melt(updateTable(table(y[,1], y[,2]), space = 0:7, N00 = TRUE)) %>% filter(Var1 + Var2 > 0)

p = ggplot(z, aes(Var1, Var2)) + theme_bw() +
    labs(x = "# of data sets\n(ESC)", 
    y = "# of data sets\n(Signature Xie)", title = "") +
    geom_tile(aes(fill = value)) + 
    geom_text(aes(label = round(value, 1))) +
    scale_fill_gradient(low = "white", high = "red")
print(p)

filter(y, ESC == 7)    

        ESC Xie
BUB1B     7   7
ECT2      7   7
MCM6      7   7
RACGAP1   7   7
TOP2A     7   7
WDR12     7   7
ZWINT     7   7

Specificity

Random signatures

PG = unique(unlist(lapply(GEX, rownames)))
set.seed(1981)
xie_random  = list(HihgRisk = sample(PG, length(xie$HihgRisk)), LowRisk = sample(PG, length(xie$LowRisk)))

set.seed(1981)
ES_random   = list(ES = sample(PG, length(gs$ES)), PRC2 = sample(PG, length(gs$PRC2)))

Random Master regulators

m1_random = mrMulti(GEX, GRN, xie_random, nsample = 50)
write.table(m1_random, file = "out/LungMR_Xie_random.txt", row.names = FALSE, col.names = TRUE, sep = "\t")

m2_random = mrMulti(GEX, GRN, ES_random,  nsample = 50)
write.table(m2_random, file = "out/LungMR_ESC_random.txt", row.names = FALSE, col.names = TRUE, sep = "\t")

Comparison of random signatures

m1_random = read.table(file = "out/LungMR_Xie_random.txt", row.names = NULL, header = TRUE, sep = "\t")
m1_random$signature = "Xie_random"
m2_random = read.table(file = "out/LungMR_ESC_random.txt", row.names = NULL, header = TRUE, sep = "\t")
m2_random$signature = "ESC_random"
mergedT = rbind(m1_random, m2_random)

uTable = filter(mergedT, p.value < 0.01, NES > 0)
dTable = filter(mergedT, p.value < 0.01, NES < 0)

x = table(uTable$Regulon, uTable$signature)
y = as.data.frame.matrix(x)
z = melt(updateTable(table(y[,1], y[,2]), space = 0:7, N00 = TRUE)) %>% filter(Var1 + Var2 > 0)

p = ggplot(z, aes(Var1, Var2)) + theme_bw() +
  labs(x = "# of data sets\n(Random signature 1)", 
    y = "# of data sets\n(Random signature 2)", title = "") +
    geom_tile(aes(fill = value)) + 
    geom_text(aes(label = round(value, 1))) +
    scale_fill_gradient(low = "white", high = "red")
print(p)

Shanghai Institute of Biochemistry and Cell Biology, Center for Excellence in Molecular Cell Science, CAS