diff --git "a/data/dataset_Receptor.csv" "b/data/dataset_Receptor.csv"
new file mode 100644--- /dev/null
+++ "b/data/dataset_Receptor.csv"
@@ -0,0 +1,3213 @@
+"keyword","repo_name","file_path","file_extension","file_size","line_count","content","language"
+"Receptor","yifanzhou330/Luminal-SNF","Fig2_and_related_ED_Fig/Metabolite_analysis.R",".R","52126","1278","#-----------------------------------------------------------------#
+# Fig3 
+#-----------------------------------------------------------------#
+rm(list = ls()) ; graphics.off()
+options(stringsAsFactors = F)
+set.seed(123)
+
+install.packages(""eoffice"") 
+library(Rtsne)
+#library(ggpubr)
+library(ggplot2)
+library(dplyr)
+
+####Load data
+load(""CBCGA_HRposHER2neg351_WES_RNAseq_CNV_Metab_Protein_20220829.Rdata"")
+load(""Fig5_S5.Rdata"")
+
+length(unique(metabolic_gene_list$b)) #4134
+# write.csv(metabolic_gene_list,""metabolic_gene_list.csv"")
+mypro <- intersect(unique(metabolic_gene_list$b),rownames(protein_log2)) #2252
+
+#####################################################################################################################
+#####################################################################################################################
+####------------------Fig 5C-D--------------####
+#####################################################################################################################
+#####################################################################################################################
+## 1. Data filtering and cleaning
+#####################################################################################################################
+
+lumimal_SNF <- as.data.frame(cbind(luminal$PatientCode,SNF_Cluster)) 
+colnames(lumimal_SNF) <- c(""PatientCode"",""SNF_subtype"")
+
+SNF1_ID <- lumimal_SNF[lumimal_SNF$SNF_subtype == ""1"",1] # 86
+SNF2_ID <- lumimal_SNF[lumimal_SNF$SNF_subtype == ""2"",1] # 89
+SNF3_ID <- lumimal_SNF[lumimal_SNF$SNF_subtype == ""3"",1] # 118
+SNF4_ID <- lumimal_SNF[lumimal_SNF$SNF_subtype == ""4"",1] # 58
+
+SNF1_PRO <- as.data.frame(protein_log2[,intersect(colnames(protein_log2),SNF1_ID)]) # 40
+SNF2_PRO <- as.data.frame(protein_log2[,intersect(colnames(protein_log2),SNF2_ID)]) # 46
+SNF3_PRO <- as.data.frame(protein_log2[,intersect(colnames(protein_log2),SNF3_ID)]) # 64
+SNF4_PRO <- as.data.frame(protein_log2[,intersect(colnames(protein_log2),SNF4_ID)]) # 29
+PT_PRO <- protein_PT_log2 # 48
+
+SNF1_PRO <- SNF1_PRO[mypro,] # 40
+SNF2_PRO <- SNF2_PRO[mypro,] # 46
+SNF3_PRO <- SNF3_PRO[mypro,] # 64
+SNF4_PRO <- SNF4_PRO[mypro,] # 29
+
+TT_PRO <- cbind(SNF1_PRO,SNF2_PRO,SNF3_PRO,SNF4_PRO) # 179
+PT_PRO <- PT_PRO[mypro,] # 48
+
+#rna
+exp.fpkm.TT_meta <- exp.fpkm.TT[rownames(exp.fpkm.TT)%in%KEGG_Meta_signature$Genes,]
+
+SNF1_rna <- exp.fpkm.TT_meta[,intersect(colnames(exp.fpkm.TT_meta),SNF1_ID)] #86
+SNF2_rna <- exp.fpkm.TT_meta[,intersect(colnames(exp.fpkm.TT_meta),SNF2_ID)] #89
+SNF3_rna <- exp.fpkm.TT_meta[,intersect(colnames(exp.fpkm.TT_meta),SNF3_ID)] #118
+SNF4_rna <- exp.fpkm.TT_meta[,intersect(colnames(exp.fpkm.TT_meta),SNF4_ID)] #58
+
+TT_rna <- cbind(SNF1_rna,SNF2_rna,SNF3_rna,SNF4_rna)
+PT_rna <- exp.fpkm.PT
+
+#pol
+SNF1_pol <- polar_metabolite_TT_MS2_log2[,intersect(colnames(polar_metabolite_TT_MS2_log2),SNF1_ID)] # 86
+SNF2_pol <- polar_metabolite_TT_MS2_log2[,intersect(colnames(polar_metabolite_TT_MS2_log2),SNF2_ID)] # 89
+SNF3_pol <- polar_metabolite_TT_MS2_log2[,intersect(colnames(polar_metabolite_TT_MS2_log2),SNF3_ID)] # 118
+SNF4_pol <- polar_metabolite_TT_MS2_log2[,intersect(colnames(polar_metabolite_TT_MS2_log2),SNF4_ID)] # 58
+TT_pol <- cbind(SNF1_pol,SNF2_pol,SNF3_pol,SNF4_pol) # 351
+PT_pol <- polar_metabolite_PT_MS2_log2 # 28
+write.csv(TT_pol,""TT_pol.CSV"")
+write.csv(PT_pol,""PT_pol.CSV"")
+
+
+SNF1_lip <- lipid_TT_MS2_log2[,intersect(colnames(lipid_TT_MS2_log2),SNF1_ID)] # 86
+SNF2_lip <- lipid_TT_MS2_log2[,intersect(colnames(lipid_TT_MS2_log2),SNF2_ID)] # 89
+SNF3_lip <- lipid_TT_MS2_log2[,intersect(colnames(lipid_TT_MS2_log2),SNF3_ID)] # 118
+SNF4_lip <- lipid_TT_MS2_log2[,intersect(colnames(lipid_TT_MS2_log2),SNF4_ID)] # 58
+
+TT_lip <- cbind(SNF1_lip,SNF2_lip,SNF3_lip,SNF4_lip) # 351
+PT_lip <- lipid_PT_MS2_log2 # 28
+# save(TT_PRO,PT_PRO,TT_pol,PT_pol,TT_lip,PT_lip,file = ""TT_PT_DATA.Rdata"")
+
+
+#####################################################################################################################
+## 2. Metabolic protein and polar metabolite subtype-specific analysis
+#####################################################################################################################
+
+##### metabolic protein #####
+comparison_matrix <- matrix(ncol=13,nrow=length(mypro))
+colnames(comparison_matrix) <- c(""SNF1"",""SNF2"",""SNF3"",""SNF4"",""max_min"",
+                                 ""SD"",""P"",""FDR"",""Tumor"",""PT"",""T_PT"",""P_TN"",""FDR_TN"")
+rownames(comparison_matrix) <- mypro
+
+for (i in rownames(comparison_matrix)){
+  comparison_matrix[i,""SNF1""] <- mean(as.numeric(SNF1_PRO[i,]),na.rm = T)
+  comparison_matrix[i,""SNF2""] <- mean(as.numeric(SNF2_PRO[i,]),na.rm = T)
+  comparison_matrix[i,""SNF3""] <- mean(as.numeric(SNF3_PRO[i,]),na.rm = T)
+  comparison_matrix[i,""SNF4""] <- mean(as.numeric(SNF4_PRO[i,]),na.rm = T)
+  comparison_matrix[i,""max_min""] <- max(as.numeric(comparison_matrix[i,1:4]))-min(as.numeric(comparison_matrix[i,1:4]))
+  comparison_matrix[i,""SD""] <- sd(TT_PRO[i,],na.rm = T)
+  a <- kruskal.test(list(as.numeric(SNF1_PRO[i,]),as.numeric(SNF2_PRO[i,]),as.numeric(SNF3_PRO[i,]),
+                         as.numeric(SNF4_PRO[i,])))
+  comparison_matrix[i,""P""] <- a$p.value
+  comparison_matrix[i,""Tumor""] <- mean(as.numeric(TT_PRO[i,]),na.rm = T)
+  comparison_matrix[i,""PT""] <- mean(as.numeric(PT_PRO[i,]),na.rm = T)
+  comparison_matrix[i,""T_PT""] <- as.numeric(comparison_matrix[i,""Tumor""])-as.numeric(comparison_matrix[i,""PT""])
+  b <- wilcox.test(as.numeric(TT_PRO[i,]),as.numeric(PT_PRO[i,]))
+  comparison_matrix[i,""P_TN""] <- b$p.value
+}
+comparison_matrix[,""FDR""] <- p.adjust(comparison_matrix[,""P""],method=""fdr"")
+comparison_matrix[,""FDR_TN""] <- p.adjust(comparison_matrix[,""P_TN""],method=""fdr"")
+comparison_matrix <- as.data.frame(comparison_matrix,stringsAsFactors = F)
+luminal_PRO_SNF <- comparison_matrix
+# write.csv(luminal_PRO_SNF,file=""./results/luminal_PRO_SNF.csv"")
+
+##### rna #####
+comparison_matrix <- matrix(ncol=13,nrow=nrow(TT_rna))
+colnames(comparison_matrix) <- c(""SNF1"",""SNF2"",""SNF3"",""SNF4"",""max_min"",
+                                 ""SD"",""P"",""FDR"",""Tumor"",""PT"",""T_PT"",""P_TN"",""FDR_TN"")
+rownames(comparison_matrix) <- rownames(TT_rna)
+
+for (i in rownames(comparison_matrix)){
+  comparison_matrix[i,""SNF1""] <- mean(as.numeric(SNF1_rna[i,]),na.rm = T)
+  comparison_matrix[i,""SNF2""] <- mean(as.numeric(SNF2_rna[i,]),na.rm = T)
+  comparison_matrix[i,""SNF3""] <- mean(as.numeric(SNF3_rna[i,]),na.rm = T)
+  comparison_matrix[i,""SNF4""] <- mean(as.numeric(SNF4_rna[i,]),na.rm = T)
+  comparison_matrix[i,""max_min""] <- max(as.numeric(comparison_matrix[i,1:4]))-min(as.numeric(comparison_matrix[i,1:4]))
+  comparison_matrix[i,""SD""] <- sd(TT_rna[i,],na.rm = T)
+  a <- kruskal.test(list(as.numeric(SNF1_rna[i,]),as.numeric(SNF2_rna[i,]),as.numeric(SNF3_rna[i,]),
+                         as.numeric(SNF4_rna[i,])))
+  comparison_matrix[i,""P""] <- a$p.value
+  comparison_matrix[i,""Tumor""] <- mean(as.numeric(TT_rna[i,]),na.rm = T)
+  comparison_matrix[i,""PT""] <- mean(as.numeric(PT_rna[i,]),na.rm = T)
+  comparison_matrix[i,""T_PT""] <- as.numeric(comparison_matrix[i,""Tumor""])-as.numeric(comparison_matrix[i,""PT""])
+  b <- wilcox.test(as.numeric(TT_rna[i,]),as.numeric(PT_rna[i,]))
+  comparison_matrix[i,""P_TN""] <- b$p.value
+}
+comparison_matrix[,""FDR""] <- p.adjust(comparison_matrix[,""P""],method=""fdr"")
+comparison_matrix[,""FDR_TN""] <- p.adjust(comparison_matrix[,""P_TN""],method=""fdr"")
+comparison_matrix <- as.data.frame(comparison_matrix,stringsAsFactors = F)
+luminal_rna_SNF <- comparison_matrix
+
+#filter:T_PT>0
+luminal_rna_SNF <- luminal_rna_SNF[luminal_rna_SNF$T_PT>0,]
+
+# write.csv(luminal_rna_SNF,file=""./results/luminal_rna_SNF.csv"")
+
+
+##### Polar metabolite #####
+comparison_matrix <- matrix(ncol=13,nrow=nrow(polar_metabolite_TT_MS2_log2))
+colnames(comparison_matrix) <- c(""SNF1"",""SNF2"",""SNF3"",""SNF4"",""max_min"",
+                                 ""SD"",""P"",""FDR"",""Tumor"",""PT"",""T_PT"",""P_TN"",""FDR_TN"")
+rownames(comparison_matrix) <- rownames(polar_metabolite_TT_MS2_log2)
+
+for (i in rownames(comparison_matrix)){
+  comparison_matrix[i,""SNF1""] <- mean(as.numeric(SNF1_pol[i,]),na.rm = T)
+  comparison_matrix[i,""SNF2""] <- mean(as.numeric(SNF2_pol[i,]),na.rm = T)
+  comparison_matrix[i,""SNF3""] <- mean(as.numeric(SNF3_pol[i,]),na.rm = T)
+  comparison_matrix[i,""SNF4""] <- mean(as.numeric(SNF4_pol[i,]),na.rm = T)
+  comparison_matrix[i,""max_min""] <- max(as.numeric(comparison_matrix[i,1:4]))-min(as.numeric(comparison_matrix[i,1:4]))
+  comparison_matrix[i,""SD""] <- sd(TT_pol[i,],na.rm = T)
+  a <- kruskal.test(list(as.numeric(SNF1_pol[i,]),as.numeric(SNF2_pol[i,]),as.numeric(SNF3_pol[i,]),
+                         as.numeric(SNF4_pol[i,])))
+  comparison_matrix[i,""P""] <- a$p.value
+  comparison_matrix[i,""Tumor""] <- mean(as.numeric(TT_pol[i,]),na.rm = T)
+  comparison_matrix[i,""PT""] <- mean(as.numeric(PT_pol[i,]),na.rm = T)
+  comparison_matrix[i,""T_PT""] <- as.numeric(comparison_matrix[i,""Tumor""])-as.numeric(comparison_matrix[i,""PT""])
+  b <- wilcox.test(as.numeric(TT_pol[i,]),as.numeric(PT_pol[i,]))
+  comparison_matrix[i,""P_TN""] <- b$p.value
+}
+comparison_matrix[,""FDR""] <- p.adjust(comparison_matrix[,""P""],method=""fdr"")
+comparison_matrix[,""FDR_TN""] <- p.adjust(comparison_matrix[,""P_TN""],method=""fdr"")
+comparison_matrix <- as.data.frame(comparison_matrix,stringsAsFactors = F)
+luminal_Pol_SNF <- comparison_matrix
+write.csv(luminal_Pol_SNF,file=""./results/luminal_Pol_SNF.csv"")
+
+##### lip #####
+comparison_matrix <- matrix(ncol=13,nrow=nrow(lipid_TT_MS2_log2))
+colnames(comparison_matrix) <- c(""SNF1"",""SNF2"",""SNF3"",""SNF4"",""max_min"",
+                                 ""SD"",""P"",""FDR"",""Tumor"",""PT"",""T_PT"",""P_TN"",""FDR_TN"")
+rownames(comparison_matrix) <- rownames(lipid_TT_MS2_log2)
+
+for (i in rownames(comparison_matrix)){
+  comparison_matrix[i,""SNF1""] <- mean(as.numeric(SNF1_lip[i,]),na.rm = T)
+  comparison_matrix[i,""SNF2""] <- mean(as.numeric(SNF2_lip[i,]),na.rm = T)
+  comparison_matrix[i,""SNF3""] <- mean(as.numeric(SNF3_lip[i,]),na.rm = T)
+  comparison_matrix[i,""SNF4""] <- mean(as.numeric(SNF4_lip[i,]),na.rm = T)
+  comparison_matrix[i,""max_min""] <- max(as.numeric(comparison_matrix[i,1:4]))-min(as.numeric(comparison_matrix[i,1:4]))
+  comparison_matrix[i,""SD""] <- sd(TT_lip[i,],na.rm = T)
+  a <- kruskal.test(list(as.numeric(SNF1_lip[i,]),as.numeric(SNF2_lip[i,]),as.numeric(SNF3_lip[i,]),
+                         as.numeric(SNF4_lip[i,])))
+  comparison_matrix[i,""P""] <- a$p.value
+  comparison_matrix[i,""Tumor""] <- mean(as.numeric(TT_lip[i,]),na.rm = T)
+  comparison_matrix[i,""PT""] <- mean(as.numeric(PT_lip[i,]),na.rm = T)
+  comparison_matrix[i,""T_PT""] <- as.numeric(comparison_matrix[i,""Tumor""])-as.numeric(comparison_matrix[i,""PT""])
+  b <- wilcox.test(as.numeric(TT_lip[i,]),as.numeric(PT_lip[i,]))
+  comparison_matrix[i,""P_TN""] <- b$p.value
+}
+comparison_matrix[,""FDR""] <- p.adjust(comparison_matrix[,""P""],method=""fdr"")
+comparison_matrix[,""FDR_TN""] <- p.adjust(comparison_matrix[,""P_TN""],method=""fdr"")
+comparison_matrix <- as.data.frame(comparison_matrix,stringsAsFactors = F)
+luminal_lip_SNF <- comparison_matrix
+# write.csv(luminal_lip_SNF,file=""./results/luminal_lip_SNF.csv"")
+
+##### Lipid_cat #####
+load(""CBCGA.Extended_MergedData_V2.5_220722.Rdata"")
+lipid_cat <- unique(CBCGA_lip_anno$Subclass)
+
+Cus_lipid_cat_TT <- matrix(ncol=ncol(TT_lip),nrow=length(lipid_cat))
+rownames(Cus_lipid_cat_TT) <- lipid_cat
+colnames(Cus_lipid_cat_TT) <- colnames(TT_lip)
+
+Cus_lipid_cat_PT <- matrix(ncol=ncol(PT_lip),nrow=length(lipid_cat))
+rownames(Cus_lipid_cat_PT) <- lipid_cat
+colnames(Cus_lipid_cat_PT) <- colnames(PT_lip)
+
+for (i in rownames(Cus_lipid_cat_TT)){
+  peak <- rownames(CBCGA_lip_anno)[CBCGA_lip_anno$Subclass==i]
+  Cus_lipid_cat_TT[i,] <- c(apply(TT_lip[peak,],2,mean))
+}
+
+for (i in rownames(Cus_lipid_cat_PT)){
+  peak <- rownames(CBCGA_lip_anno)[CBCGA_lip_anno$Subclass==i]
+  Cus_lipid_cat_PT[i,] <- c(apply(PT_lip[peak,],2,mean))
+}
+
+SNF1_lip_cat <- Cus_lipid_cat_TT[,intersect(colnames(lipid_TT_MS2_log2),SNF1_ID)] # 86
+SNF2_lip_cat <- Cus_lipid_cat_TT[,intersect(colnames(lipid_TT_MS2_log2),SNF2_ID)] # 89
+SNF3_lip_cat <- Cus_lipid_cat_TT[,intersect(colnames(lipid_TT_MS2_log2),SNF3_ID)] # 118
+SNF4_lip_cat <- Cus_lipid_cat_TT[,intersect(colnames(lipid_TT_MS2_log2),SNF4_ID)] # 58
+
+TT_lip_cat <- cbind(SNF1_lip_cat,SNF2_lip_cat,SNF3_lip_cat,SNF4_lip_cat) # 351
+PT_lip_cat <- Cus_lipid_cat_PT
+
+comparison_matrix <- matrix(ncol=13,nrow=nrow(TT_lip_cat))
+colnames(comparison_matrix) <- c(""SNF1"",""SNF2"",""SNF3"",""SNF4"",""max_min"",
+                                 ""SD"",""P"",""FDR"",""Tumor"",""PT"",""T_PT"",""P_TN"",""FDR_TN"")
+rownames(comparison_matrix) <- rownames(TT_lip_cat)
+
+for (i in rownames(comparison_matrix)){
+  comparison_matrix[i,""SNF1""] <- mean(as.numeric(SNF1_lip_cat[i,]),na.rm = T)
+  comparison_matrix[i,""SNF2""] <- mean(as.numeric(SNF2_lip_cat[i,]),na.rm = T)
+  comparison_matrix[i,""SNF3""] <- mean(as.numeric(SNF3_lip_cat[i,]),na.rm = T)
+  comparison_matrix[i,""SNF4""] <- mean(as.numeric(SNF4_lip_cat[i,]),na.rm = T)
+  comparison_matrix[i,""max_min""] <- max(as.numeric(comparison_matrix[i,1:4]))-min(as.numeric(comparison_matrix[i,1:4]))
+  comparison_matrix[i,""SD""] <- sd(TT_lip_cat[i,],na.rm = T)
+  a <- kruskal.test(list(as.numeric(SNF1_lip_cat[i,]),as.numeric(SNF2_lip_cat[i,]),as.numeric(SNF3_lip_cat[i,]),
+                         as.numeric(SNF4_lip_cat[i,])))
+  comparison_matrix[i,""P""] <- a$p.value
+  comparison_matrix[i,""Tumor""] <- mean(as.numeric(TT_lip_cat[i,]),na.rm = T)
+  comparison_matrix[i,""PT""] <- mean(as.numeric(PT_lip_cat[i,]),na.rm = T)
+  comparison_matrix[i,""T_PT""] <- as.numeric(comparison_matrix[i,""Tumor""])-as.numeric(comparison_matrix[i,""PT""])
+  b <- wilcox.test(as.numeric(TT_lip_cat[i,]),as.numeric(PT_lip_cat[i,]))
+  comparison_matrix[i,""P_TN""] <- b$p.value
+}
+comparison_matrix[,""FDR""] <- p.adjust(comparison_matrix[,""P""],method=""fdr"")
+comparison_matrix[,""FDR_TN""] <- p.adjust(comparison_matrix[,""P_TN""],method=""fdr"")
+comparison_matrix <- as.data.frame(comparison_matrix,stringsAsFactors = F)
+luminal_lip_cat_SNF <- comparison_matrix
+# write.csv(luminal_lip_cat_SNF,file=""./results/luminal_lip_cat_SNF.csv"")
+
+##### Scale #####
+scale_PRO<-t(scale(t(protein_log2),center=T,scale=T))
+scale_Pol<-t(scale(t(polar_metabolite_TT_MS2_log2),center=T,scale=T))
+scale_lip_cat<-t(scale(t(Cus_lipid_cat_TT),center=T,scale=T))
+scale_rna<-t(scale(t(exp.fpkm.TT_meta),center=T,scale=T))
+scale_rna <- scale_rna[rownames(scale_rna)%in%rownames(luminal_rna_SNF),]
+
+
+#PRO
+SNF1_matrix <- scale_PRO[,intersect(colnames(scale_PRO),SNF1_ID)]
+SNF2_matrix <- scale_PRO[,intersect(colnames(scale_PRO),SNF2_ID)]
+SNF3_matrix <- scale_PRO[,intersect(colnames(scale_PRO),SNF3_ID)]
+SNF4_matrix <- scale_PRO[,intersect(colnames(scale_PRO),SNF4_ID)]
+
+comparison_matrix <- matrix(ncol=4,nrow=nrow(scale_PRO))
+colnames(comparison_matrix) <- c(""SNF1"",""SNF2"",""SNF3"",""SNF4"")
+rownames(comparison_matrix) <- rownames(scale_PRO)
+
+for (i in rownames(comparison_matrix)){
+  comparison_matrix[i,""SNF1""] <- mean(as.numeric(SNF1_matrix[i,]),na.rm = T)
+  comparison_matrix[i,""SNF2""] <- mean(as.numeric(SNF2_matrix[i,]),na.rm = T)
+  comparison_matrix[i,""SNF3""] <- mean(as.numeric(SNF3_matrix[i,]),na.rm = T)
+  comparison_matrix[i,""SNF4""] <- mean(as.numeric(SNF4_matrix[i,]),na.rm = T)
+}
+comparison_matrix<-as.data.frame(comparison_matrix,stringsAsFactors = F)
+scale_pro_SNF<-comparison_matrix
+# write.table(comparison_matrix,file=""scale_pro_PAM50.txt"",sep=""\t"")
+
+#rna
+SNF1_matrix <- scale_rna[,intersect(colnames(scale_rna),SNF1_ID)]
+SNF2_matrix <- scale_rna[,intersect(colnames(scale_rna),SNF2_ID)]
+SNF3_matrix <- scale_rna[,intersect(colnames(scale_rna),SNF3_ID)]
+SNF4_matrix <- scale_rna[,intersect(colnames(scale_rna),SNF4_ID)]
+
+comparison_matrix <- matrix(ncol=4,nrow=nrow(scale_rna))
+colnames(comparison_matrix) <- c(""SNF1"",""SNF2"",""SNF3"",""SNF4"")
+rownames(comparison_matrix) <- rownames(scale_rna)
+
+for (i in rownames(comparison_matrix)){
+  comparison_matrix[i,""SNF1""] <- mean(as.numeric(SNF1_matrix[i,]),na.rm = T)
+  comparison_matrix[i,""SNF2""] <- mean(as.numeric(SNF2_matrix[i,]),na.rm = T)
+  comparison_matrix[i,""SNF3""] <- mean(as.numeric(SNF3_matrix[i,]),na.rm = T)
+  comparison_matrix[i,""SNF4""] <- mean(as.numeric(SNF4_matrix[i,]),na.rm = T)
+}
+comparison_matrix<-as.data.frame(comparison_matrix,stringsAsFactors = F)
+scale_rna_SNF<-comparison_matrix
+# write.table(comparison_matrix,file=""scale_rna_PAM50.txt"",sep=""\t"")
+
+#POL
+SNF1_matrix <- scale_Pol[,intersect(colnames(scale_Pol),SNF1_ID)]
+SNF2_matrix <- scale_Pol[,intersect(colnames(scale_Pol),SNF2_ID)]
+SNF3_matrix <- scale_Pol[,intersect(colnames(scale_Pol),SNF3_ID)]
+SNF4_matrix <- scale_Pol[,intersect(colnames(scale_Pol),SNF4_ID)]
+
+comparison_matrix <- matrix(ncol=4,nrow=nrow(scale_Pol))
+colnames(comparison_matrix) <- c(""SNF1"",""SNF2"",""SNF3"",""SNF4"")
+rownames(comparison_matrix) <- rownames(scale_Pol)
+
+for (i in rownames(comparison_matrix)){
+  comparison_matrix[i,""SNF1""] <- mean(as.numeric(SNF1_matrix[i,]),na.rm = T)
+  comparison_matrix[i,""SNF2""] <- mean(as.numeric(SNF2_matrix[i,]),na.rm = T)
+  comparison_matrix[i,""SNF3""] <- mean(as.numeric(SNF3_matrix[i,]),na.rm = T)
+  comparison_matrix[i,""SNF4""] <- mean(as.numeric(SNF4_matrix[i,]),na.rm = T)
+}
+comparison_matrix<-as.data.frame(comparison_matrix,stringsAsFactors = F)
+scale_Pol_SNF<-comparison_matrix
+
+#lip
+SNF1_matrix <- scale_lip_cat[,intersect(colnames(scale_lip_cat),SNF1_ID)]
+SNF2_matrix <- scale_lip_cat[,intersect(colnames(scale_lip_cat),SNF2_ID)]
+SNF3_matrix <- scale_lip_cat[,intersect(colnames(scale_lip_cat),SNF3_ID)]
+SNF4_matrix <- scale_lip_cat[,intersect(colnames(scale_lip_cat),SNF4_ID)]
+
+comparison_matrix <- matrix(ncol=4,nrow=nrow(scale_lip_cat))
+colnames(comparison_matrix) <- c(""SNF1"",""SNF2"",""SNF3"",""SNF4"")
+rownames(comparison_matrix) <- rownames(scale_lip_cat)
+
+for (i in rownames(comparison_matrix)){
+  comparison_matrix[i,""SNF1""] <- mean(as.numeric(SNF1_matrix[i,]),na.rm = T)
+  comparison_matrix[i,""SNF2""] <- mean(as.numeric(SNF2_matrix[i,]),na.rm = T)
+  comparison_matrix[i,""SNF3""] <- mean(as.numeric(SNF3_matrix[i,]),na.rm = T)
+  comparison_matrix[i,""SNF4""] <- mean(as.numeric(SNF4_matrix[i,]),na.rm = T)
+}
+comparison_matrix<-as.data.frame(comparison_matrix,stringsAsFactors = F)
+scale_lip_cat_SNF<-comparison_matrix
+
+
+# node_scale_PAM50<-rbind(scale_pro_PAM50,scale_pol_PAM50)
+
+
+#####################################################################################################################
+## 3. Metabolic protein and metabolite correlation network construction
+#####################################################################################################################
+
+# It takes several time to finish part 3/4 and their figures are not drawn by R....... temporarily skip them 
+
+## Metabolic protein correlation network construction
+pro_Cor.Res <- matrix(nrow=nrow(TT_PRO),ncol=nrow(TT_PRO))
+pro_P.val <- matrix(nrow=nrow(TT_PRO),ncol=nrow(TT_PRO))
+colnames(pro_Cor.Res) <- rownames(TT_PRO)
+rownames(pro_Cor.Res) <- rownames(TT_PRO)
+colnames(pro_P.val) <- rownames(TT_PRO)
+rownames(pro_P.val) <- rownames(TT_PRO)
+
+TT_PRO <- t(TT_PRO)
+
+pb <- txtProgressBar(style=3)
+for (i in 1:ncol(TT_PRO)){
+  for (j in 1:ncol(TT_PRO)){
+    TEMP_Res <- cor.test(TT_PRO[,i],TT_PRO[,j], method = ""spearman"")
+    pro_Cor.Res[i,j] <- TEMP_Res$estimate
+    pro_P.val[i,j] <- TEMP_Res$p.value
+  }
+  setTxtProgressBar(pb, i/ncol(TT_PRO))
+}
+
+pro_FDR <- matrix(p.adjust(pro_P.val,method=""fdr""),ncol=2252)
+colnames(pro_FDR) <- colnames(pro_P.val)
+rownames(pro_FDR) <- rownames(pro_P.val)
+
+## rna : Metabolic transcriptomics correlation network construction
+TT_rna <- TT_rna[rownames(TT_rna)%in%rownames(luminal_rna_SNF),]
+rna_Cor.Res <- matrix(nrow=nrow(TT_rna),ncol=nrow(TT_rna))
+rna_P.val <- matrix(nrow=nrow(TT_rna),ncol=nrow(TT_rna))
+colnames(rna_Cor.Res) <- rownames(TT_rna)
+rownames(rna_Cor.Res) <- rownames(TT_rna)
+colnames(rna_P.val) <- rownames(TT_rna)
+rownames(rna_P.val) <- rownames(TT_rna)
+
+TT_rna <- t(TT_rna)
+
+pb <- txtProgressBar(style=3)
+for (i in 1:ncol(TT_rna)){
+  for (j in 1:ncol(TT_rna)){
+    TEMP_Res <- cor.test(TT_rna[,i],TT_rna[,j], method = ""spearman"")
+    rna_Cor.Res[i,j] <- TEMP_Res$estimate
+    rna_P.val[i,j] <- TEMP_Res$p.value
+  }
+  setTxtProgressBar(pb, i/ncol(TT_rna))
+}
+
+rna_FDR <- matrix(p.adjust(rna_P.val,method=""fdr""),ncol=836)
+colnames(rna_FDR) <- colnames(rna_P.val)
+rownames(rna_FDR) <- rownames(rna_P.val)
+
+# save(rna_Cor.Res,rna_P.val,rna_FDR,file=""rna_Cor_spearman.Rdata"")
+
+##lipid correlation network construction
+lip_Cor.Res <- matrix(nrow=nrow(TT_lip_cat),ncol=nrow(TT_lip_cat))
+lip_P.val <- matrix(nrow=nrow(TT_lip_cat),ncol=nrow(TT_lip_cat))
+colnames(lip_Cor.Res) <- rownames(TT_lip_cat)
+rownames(lip_Cor.Res) <- rownames(TT_lip_cat)
+colnames(lip_P.val) <- rownames(TT_lip_cat)
+rownames(lip_P.val) <- rownames(TT_lip_cat)
+
+TT_lip_cat <- t(TT_lip_cat)
+
+pb <- txtProgressBar(style=3)
+for (i in 1:ncol(TT_lip_cat)){
+  for (j in 1:ncol(TT_lip_cat)){
+    TEMP_Res <- cor.test(TT_lip_cat[,i],TT_lip_cat[,j], method = ""spearman"")
+    lip_Cor.Res[i,j] <- TEMP_Res$estimate
+    lip_P.val[i,j] <- TEMP_Res$p.value
+  }
+  setTxtProgressBar(pb, i/ncol(TT_lip_cat))
+}
+
+lip_FDR <- matrix(p.adjust(lip_P.val,method=""fdr""),ncol=46)
+colnames(lip_FDR) <- colnames(lip_P.val)
+rownames(lip_FDR) <- rownames(lip_P.val)
+
+# save(lip_Cor.Res,lip_P.val,lip_FDR,file=""lip_Cor_spearman.Rdata"")
+
+
+## Polar metabolite correlation network construction
+pol_Cor.Res <- matrix(nrow=nrow(TT_pol),ncol=nrow(TT_pol))
+pol_P.val <- matrix(nrow=nrow(TT_pol),ncol=nrow(TT_pol))
+colnames(pol_Cor.Res) <- rownames(TT_pol)
+rownames(pol_Cor.Res) <- rownames(TT_pol)
+colnames(pol_P.val) <- rownames(TT_pol)
+rownames(pol_P.val) <- rownames(TT_pol)
+
+TT_pol <- t(TT_pol)
+
+pb <- txtProgressBar(style=3)
+for (i in 1:ncol(TT_pol)){
+  for (j in 1:ncol(TT_pol)){
+    TEMP_Res <- cor.test(TT_pol[,i],TT_pol[,j], method = ""spearman"")
+    pol_Cor.Res[i,j] <- TEMP_Res$estimate
+    pol_P.val[i,j] <- TEMP_Res$p.value
+  }
+  setTxtProgressBar(pb, i/ncol(TT_pol))
+}
+
+pol_FDR <- matrix(p.adjust(pol_P.val,method=""fdr""),ncol=669)
+colnames(pol_FDR) <- colnames(pol_P.val)
+rownames(pol_FDR) <- rownames(pol_P.val)
+
+# save(pol_Cor.Res,pol_P.val,pol_FDR,file=""pol_Cor_spearman.Rdata"")
+
+
+## Metabolic protein and Polar metabolite correlation network construction
+Luminal_PRO_POL_ID <- intersect(rownames(TT_PRO),rownames(TT_pol)) # 179
+
+pro_pol_matrix <- cbind(TT_PRO[Luminal_PRO_POL_ID,],TT_pol[Luminal_PRO_POL_ID,]) #179*2921
+dim(pro_pol_matrix)
+
+pro_pol_Cor.Res <- matrix(nrow=ncol(pro_pol_matrix),ncol=ncol(pro_pol_matrix))
+pro_pol_P.val <- matrix(nrow=ncol(pro_pol_matrix),ncol=ncol(pro_pol_matrix))
+colnames(pro_pol_Cor.Res) <- colnames(pro_pol_matrix)
+rownames(pro_pol_Cor.Res) <- colnames(pro_pol_matrix)
+colnames(pro_pol_P.val) <- colnames(pro_pol_matrix)
+rownames(pro_pol_P.val) <- colnames(pro_pol_matrix)
+
+pb <- txtProgressBar(style=3)
+for (i in 1:ncol(pro_pol_matrix)){
+  for (j in 1:ncol(pro_pol_matrix)){
+    TEMP_Res <- cor.test(pro_pol_matrix[,i],pro_pol_matrix[,j], method = ""spearman"")
+    pro_pol_Cor.Res[i,j] <- TEMP_Res$estimate
+    pro_pol_P.val[i,j] <- TEMP_Res$p.value
+  }
+  setTxtProgressBar(pb, i/ncol(pro_pol_matrix))
+}
+
+pro_pol_FDR <- matrix(p.adjust(pro_pol_P.val,method=""fdr""),ncol=2921)
+colnames(pro_pol_FDR) <- colnames(pro_pol_P.val)
+rownames(pro_pol_FDR) <- rownames(pro_pol_P.val)
+
+save(pro_pol_Cor.Res,pro_pol_P.val,pro_pol_FDR,file=""pro_pol_Cor_spearman.Rdata"")
+write.csv(pro_pol_Cor.Res,""pro_pol_Cor.Res.csv"")
+pro_pol_Cor.Res[""CAD"",""M133T340_POS""]
+pro_pol_P.val[""CAD"",""M133T340_POS""]
+pro_pol_FDR[""CAD"",""M133T340_POS""]
+
+##### [edge] #####
+##### Metabolic protein [edge] #####
+res.matrix_pro<-data.frame(""a"",""b"",0.5,0.05)
+colnames(res.matrix_pro)<-c(""source"",""target"",""weight"",""FDR"")
+
+for (i in 1:2252){
+  for (j in (i+1):2252){
+    if ((j!=i) & (pro_pol_FDR[i,j]<0.05) & (pro_pol_Cor.Res[i,j] > 0.4)){
+      res.matrix_pro<-rbind(res.matrix_pro,c(rownames(pro_pol_Cor.Res)[i],colnames(pro_pol_Cor.Res)[j],
+                                             pro_pol_Cor.Res[i,j],pro_pol_FDR[i,j]))
+    }
+  }
+  setTxtProgressBar(pb, i/2252)
+}
+#19554
+res.matrix_pro<-res.matrix_pro[-1,]
+res.matrix_pro$type<-rep(""undirected"",n=nrow(res.matrix_pro))
+
+#Deleting the protein without annotation
+anno_manual_pro<-read.csv(""anno_manual_pro.csv"")
+rownames(anno_manual_pro)<-anno_manual_pro$X
+anno_manual_pro<-anno_manual_pro[,-1]
+
+res.matrix_pro$anno_source<-anno_manual_pro[res.matrix_pro$source,""recon_kegg""]
+res.matrix_pro$anno_target<-anno_manual_pro[res.matrix_pro$target,""recon_kegg""]
+res.matrix_pro$anno_source[which(res.matrix_pro$anno_source=="""")]<-NA
+res.matrix_pro$anno_target[which(res.matrix_pro$anno_target=="""")]<-NA
+res.matrix_pro_final<-na.omit(res.matrix_pro) #4541
+write.csv(res.matrix_pro_final[,c(1,2,3,5)],""res.matrix_pro.csv"",row.names=FALSE)
+
+##### rna [edge] #####
+res.matrix_rna<-data.frame(""a"",""b"",0.5,0.05)
+colnames(res.matrix_rna)<-c(""source"",""target"",""weight"",""FDR"")
+
+load(""rna_Cor_spearman.Rdata"")
+rna_FDR[is.na(rna_FDR)] <- 1
+rna_Cor.Res[is.na(rna_Cor.Res)] <- 0
+for (i in 1:836){
+  for (j in (i+1):836){
+    if ((j!=i) & (rna_FDR[i,j]<0.05) & (rna_Cor.Res[i,j] > 0.4)){
+      res.matrix_rna<-rbind(res.matrix_rna,c(rownames(rna_Cor.Res)[i],colnames(rna_Cor.Res)[j],
+                                             rna_Cor.Res[i,j],rna_FDR[i,j]))
+    }
+  }
+  setTxtProgressBar(pb, i/836)
+}
+#4645
+res.matrix_rna<-res.matrix_rna[-1,]
+res.matrix_rna$type<-rep(""undirected"",n=nrow(res.matrix_rna))
+write.csv(res.matrix_rna[,c(1,2,3,5)],""res.matrix_rna.csv"",row.names=FALSE)
+
+##### lip [edge] #####
+res.matrix_lip<-data.frame(""a"",""b"",0.5,0.05)
+colnames(res.matrix_lip)<-c(""source"",""target"",""weight"",""FDR"")
+
+for (i in 1:46){
+  for (j in (i+1):46){
+    if ((j!=i) & (lip_FDR[i,j]<0.05) & (lip_Cor.Res[i,j] > 0.4)){
+      res.matrix_lip<-rbind(res.matrix_lip,c(rownames(lip_Cor.Res)[i],colnames(lip_Cor.Res)[j],
+                                             lip_Cor.Res[i,j],lip_FDR[i,j]))
+    }
+  }
+  setTxtProgressBar(pb, i/46)
+}
+#697
+res.matrix_lip<-res.matrix_lip[-1,]
+res.matrix_lip$type<-rep(""undirected"",n=nrow(res.matrix_lip))
+write.csv(res.matrix_lip[,c(1,2,3,5)],""res.matrix_lip.csv"",row.names=FALSE)
+
+##### Polar Metabolite [edge] #####
+res.matrix_pol<-data.frame(""a"",""b"",0.5,0.05)
+colnames(res.matrix_pol)<-c(""source"",""target"",""weight"",""FDR"")
+
+for (i in 2253:2921){
+  for (j in (i+1):2921){
+    if ((j!=i) & (pro_pol_FDR[i,j]<0.05) & (pro_pol_Cor.Res[i,j] > 0.4)){
+      res.matrix_pol<-rbind(res.matrix_pol,c(rownames(pro_pol_Cor.Res)[i],colnames(pro_pol_Cor.Res)[j],
+                                             pro_pol_Cor.Res[i,j],pro_pol_FDR[i,j]))
+    }
+  }
+  setTxtProgressBar(pb, i/669)
+}
+#17467
+res.matrix_pol<-res.matrix_pol[-1,]
+res.matrix_pol$type<-rep(""undirected"",n=nrow(res.matrix_pol))
+write.csv(res.matrix_pol[,c(1,2,3,5)],""res.matrix_pol.csv"",row.names=FALSE)
+
+
+##### [node] #####
+# scale_pro_PAM50<-read.table(""scale_pro_PAM50.txt"",sep=""\t"")
+# scale_pol_PAM50<-read.table(""scale_pol_PAM50.txt"",sep=""\t"")
+
+scale_node<-rbind(scale_pro_SNF,scale_Pol_SNF)
+
+scale_node <- scale_lip_cat_SNF
+scale_rna_SNF <- na.omit(scale_rna_SNF)
+scale_node <- scale_rna_SNF
+
+
+max(scale_node)
+min(scale_node)
+for(i in 1:ncol(scale_node)){
+  scale_node[,i][which(scale_node[,i] >= 1)]<- 1
+  scale_node[,i][which(scale_node[,i] >= 0.8 & scale_node[,i] < 1)]<- 0.8
+  scale_node[,i][which(scale_node[,i] >= 0.6 & scale_node[,i] < 0.8)]<- 0.6
+  scale_node[,i][which(scale_node[,i] >= 0.4 & scale_node[,i] < 0.6)]<- 0.4
+  scale_node[,i][which(scale_node[,i] >= 0.2 & scale_node[,i] < 0.4)]<- 0.2
+  scale_node[,i][which(scale_node[,i] >= 0 & scale_node[,i] < 0.2)]<- 0
+  scale_node[,i][which(scale_node[,i] >= -0.2 & scale_node[,i] < 0)]<- -0.2
+  scale_node[,i][which(scale_node[,i] >= -0.4 & scale_node[,i] < -0.2)]<- -0.4
+  scale_node[,i][which(scale_node[,i] >= -0.6 & scale_node[,i] < -0.4)]<- -0.6
+  scale_node[,i][which(scale_node[,i] >= -0.8 & scale_node[,i] < -0.6)]<- -0.8
+  scale_node[,i][which(scale_node[,i] < -0.8)]<- -1
+}
+
+
+anno_manual<-read.csv(""anno_manual_pro.csv"")
+rownames(anno_manual)<-anno_manual$id
+anno_manual<-anno_manual[,-1]
+node_SNF<-luminal_rna_SNF
+
+res.matrix_cus<-res.matrix_pol
+res.matrix_cus<-res.matrix_pro_final
+res.matrix_cus<-res.matrix_lip
+res.matrix_cus<-res.matrix_rna
+
+node_cor<-union(unique(res.matrix_cus$source),unique(res.matrix_cus$target))
+node_anno<-data.frame(node_cor,""type"",""Label"",""pathway_class"",
+                      ""SNF1"",""SNF2"",""SNF3"",""SNF4"",""max_min"",
+                      ""SD"",""P"",""FDR"",""Tumor"",""PT"",""T_PT"",""P_TN"",""FDR_TN"")
+colnames(node_anno)<-c(""Id"",""type"",""Label"",""pathway_class"",
+                       ""SNF1"",""SNF2"",""SNF3"",""SNF4"",""max_min"",
+                       ""SD"",""P"",""FDR"",""Tumor"",""PT"",""T_PT"",""P_TN"",""FDR_TN"")
+rownames(node_anno)<-node_anno$Id
+
+for (i in rownames(node_anno)){
+  if (i %in% rownames(CBCGA_pol_anno)){
+    node_anno[i,""type""]<-""metabolite""
+    node_anno[i,""Label""]<-CBCGA_pol_anno[i,""Putative_metabolite_name""]
+    node_anno[i,""pathway_class""]<-anno_manual[i,""subclass_final_manual""]
+    node_anno[i,""subclass""]<-CBCGA_pol_anno[i,""Metabolite_class""]
+  } else {
+    node_anno[i,""type""]<-""protein""
+    node_anno[i,""Label""]<-node_anno[i,""Id""]
+    node_anno[i,""pathway_class""]<-anno_manual_pro[i,""recon_kegg""]}
+  node_anno[i,5:8]<-scale_node[i,]
+  node_anno[i,9:17]<-node_SNF[i,5:13]
+}
+node_anno_pol<-node_anno
+node_anno_pro<-node_anno
+node_anno_lip<-node_anno
+node_anno_rna<-node_anno
+
+write.csv(node_anno,""node_anno_pro.csv"",row.names = F)
+write.csv(node_anno,""node_anno_pol.csv"",row.names = F)
+write.csv(node_anno,""node_anno_lip.csv"",row.names = F)
+write.csv(node_anno,""node_anno_rna.csv"",row.names = F)
+
+## Refer to Gephi
+
+#################lipboxplot########################
+if (F) {
+  lipid_main <- unique(CBCGA_lip_anno$Lipid.super.class)
+  
+  Cus_lipid_main_TT <- matrix(ncol=ncol(TT_lip),nrow=length(lipid_main))
+  rownames(Cus_lipid_main_TT) <- lipid_main
+  colnames(Cus_lipid_main_TT) <- colnames(TT_lip)
+  
+  for (i in rownames(Cus_lipid_main_TT)){
+    peak <- rownames(CBCGA_lip_anno)[CBCGA_lip_anno$Lipid.super.class==i]
+    Cus_lipid_main_TT[i,] <- c(apply(TT_lip[peak,],2,mean))
+  }
+  
+  SNF1_matrix <- Cus_lipid_main_TT[,intersect(colnames(Cus_lipid_main_TT),SNF1_ID)]
+  SNF2_matrix <- Cus_lipid_main_TT[,intersect(colnames(Cus_lipid_main_TT),SNF2_ID)]
+  SNF3_matrix <- Cus_lipid_main_TT[,intersect(colnames(Cus_lipid_main_TT),SNF3_ID)]
+  SNF4_matrix <- Cus_lipid_main_TT[,intersect(colnames(Cus_lipid_main_TT),SNF4_ID)]
+  
+  comparison_matrix <- matrix(ncol=4,nrow=nrow(Cus_lipid_main_TT))
+  colnames(comparison_matrix) <- c(""SNF1"",""SNF2"",""SNF3"",""SNF4"")
+  rownames(comparison_matrix) <- rownames(Cus_lipid_main_TT)
+  
+  for (i in rownames(comparison_matrix)){
+    comparison_matrix[i,""SNF1""] <- mean(as.numeric(SNF1_matrix[i,]),na.rm = T)
+    comparison_matrix[i,""SNF2""] <- mean(as.numeric(SNF2_matrix[i,]),na.rm = T)
+    comparison_matrix[i,""SNF3""] <- mean(as.numeric(SNF3_matrix[i,]),na.rm = T)
+    comparison_matrix[i,""SNF4""] <- mean(as.numeric(SNF4_matrix[i,]),na.rm = T)
+  }
+  comparison_matrix<-as.data.frame(comparison_matrix,stringsAsFactors = F)
+  Cus_lipid_main_TT_SNF<-comparison_matrix
+}
+library(tidyr)
+library(pheatmap)
+library(RColorBrewer)
+
+luminal_lip_SNF <- read.csv(""./results/luminal_lip_SNF.csv"",header = T,row.names = 1)
+luminal_lip_SNF$main <- CBCGA_lip_anno$Lipid.super.class
+luminal_lip_SNF$SNF1_PT <- luminal_lip_SNF$SNF1-luminal_lip_SNF$PT
+luminal_lip_SNF$SNF2_PT <- luminal_lip_SNF$SNF2-luminal_lip_SNF$PT
+luminal_lip_SNF$SNF3_PT <- luminal_lip_SNF$SNF3-luminal_lip_SNF$PT
+luminal_lip_SNF$SNF4_PT <- luminal_lip_SNF$SNF4-luminal_lip_SNF$PT
+
+plotdata <- luminal_lip_SNF[,c(14:18)]
+plotdata$peak <- rownames(plotdata)
+
+plotdata_FA <- subset(plotdata,plotdata$main==""Fatty acyls [FA]"")
+kruskal.test(list(plotdata_FA$SNF1_PT,
+                  plotdata_FA$SNF2_PT,
+                  plotdata_FA$SNF3_PT,
+                  plotdata_FA$SNF4_PT))#p-value = 0.0004746***
+plotdata_GL <- subset(plotdata,plotdata$main==""Glycerolipids [GL]"")
+kruskal.test(list(plotdata_GL$SNF1_PT,
+                  plotdata_GL$SNF2_PT,
+                  plotdata_GL$SNF3_PT,
+                  plotdata_GL$SNF4_PT))#p-value = 0.008005**
+plotdata_GP <- subset(plotdata,plotdata$main==""Glycerophospholipids [GP]"")
+kruskal.test(list(plotdata_GP$SNF1_PT,
+                  plotdata_GP$SNF2_PT,
+                  plotdata_GP$SNF3_PT,
+                  plotdata_GP$SNF4_PT))#p-value = 9.113e-14***
+plotdata_SP <- subset(plotdata,plotdata$main==""Sphingolipids [SP]"")
+kruskal.test(list(plotdata_SP$SNF1_PT,
+                  plotdata_SP$SNF2_PT,
+                  plotdata_SP$SNF3_PT,
+                  plotdata_SP$SNF4_PT))#0.002252**
+plotdata_ST <- subset(plotdata,plotdata$main==""Sterol Lipids [ST]"")
+kruskal.test(list(plotdata_ST$SNF1_PT,
+                  plotdata_ST$SNF2_PT,
+                  plotdata_ST$SNF3_PT,
+                  plotdata_ST$SNF4_PT))#p-value = 0.761
+
+
+
+plotdata <- gather(plotdata,SNF,abundance,-main,-peak)
+
+library(ggplot2)
+
+#please select optimal color and xlim, ylim
+
+ggplot(plotdata,aes(x = main, y = abundance, fill = SNF)) +
+  scale_fill_manual(values=c(SNF1_PT=""#3D76AE"",SNF2_PT=""#53AD4A"",SNF3_PT=""#EDAB3C"",SNF4_PT=""#C3392A"") )+
+  geom_boxplot(outlier.shape = NA,linetype=""dashed"")+
+  stat_boxplot(outlier.shape=NA,aes(ymin=..lower..,ymax=..upper..))+
+  stat_boxplot(geom = ""errorbar"",aes(ymin=..ymax..),width=0.75)+
+  stat_boxplot(geom = ""errorbar"",aes(ymax=..ymin..),width=0.75)+ 
+  labs(x=""lipid abundance"",y=""Mean |log2FC| between tumor and normal"")+
+  # ggtitle(""Boxplot of hub gene"") +
+  theme(panel.grid.major=element_blank(), 
+        panel.grid.minor=element_blank(),
+        panel.background = element_blank())+
+  theme(panel.border = element_rect(fill=NA,color=""black"", size=0.8, linetype=""solid""))
+#boxplot of mean log2FC of lipid categories
+boxplot(log2FC~catagory,data=Cus_lipid_main_TT_SNF,
+        at=c(1:5),xlim=c(0,6),ylim=c(-2,2),
+        col= c(""#3D76AE"" ,""#53AD4A"", ""#EDAB3C"", ""#C3392A""), 
+        #names=names(table(comparison_TT_PT_matrix_cat$catagory_whole)),
+        outline=F,
+        main=""Fold change""
+)
+abline(h=0,col = ""red"", lwd = 2, lty = 2)
+
+#############polar boxplot################
+
+luminal_Pol_SNF <- read.csv(""./results/luminal_Pol_SNF.csv"",header = T,row.names = 1)
+luminal_Pol_SNF$main <- CBCGA_pol_anno$Metabolite_class
+luminal_Pol_SNF$SNF1_PT <- luminal_Pol_SNF$SNF1-luminal_Pol_SNF$PT
+luminal_Pol_SNF$SNF2_PT <- luminal_Pol_SNF$SNF2-luminal_Pol_SNF$PT
+luminal_Pol_SNF$SNF3_PT <- luminal_Pol_SNF$SNF3-luminal_Pol_SNF$PT
+luminal_Pol_SNF$SNF4_PT <- luminal_Pol_SNF$SNF4-luminal_Pol_SNF$PT
+
+plotdata <- luminal_Pol_SNF[,c(14:18)]
+plotdata$peak <- rownames(plotdata)
+
+plotdata_AA <- subset(plotdata,plotdata$main==""Amino acid"")
+kruskal.test(list(plotdata_AA$SNF1_PT,
+                  plotdata_AA$SNF2_PT,
+                  plotdata_AA$SNF3_PT,
+                  plotdata_AA$SNF4_PT))#p-value = 4.952e-05***
+plotdata_carbon <- subset(plotdata,plotdata$main==""Carbohydrates"")
+kruskal.test(list(plotdata_carbon$SNF1_PT,
+                  plotdata_carbon$SNF2_PT,
+                  plotdata_carbon$SNF3_PT,
+                  plotdata_carbon$SNF4_PT))#p-value = 0.4084
+plotdata_lip <- subset(plotdata,plotdata$main==""Lipid"")
+kruskal.test(list(plotdata_lip$SNF1_PT,
+                  plotdata_lip$SNF2_PT,
+                  plotdata_lip$SNF3_PT,
+                  plotdata_lip$SNF4_PT))#p-value = 0.02426*
+plotdata_nucle <- subset(plotdata,plotdata$main==""Nucleotide"")
+kruskal.test(list(plotdata_nucle$SNF1_PT,
+                  plotdata_nucle$SNF2_PT,
+                  plotdata_nucle$SNF3_PT,
+                  plotdata_nucle$SNF4_PT))#0.3861
+plotdata_Other <- subset(plotdata,plotdata$main==""Other"")
+kruskal.test(list(plotdata_Other$SNF1_PT,
+                  plotdata_Other$SNF2_PT,
+                  plotdata_Other$SNF3_PT,
+                  plotdata_Other$SNF4_PT))#p-value = 0.6101
+plotdata_pep <- subset(plotdata,plotdata$main==""Peptide"")
+kruskal.test(list(plotdata_pep$SNF1_PT,
+                  plotdata_pep$SNF2_PT,
+                  plotdata_pep$SNF3_PT,
+                  plotdata_pep$SNF4_PT))#p-value = 0.0002545***
+plotdata_vita <- subset(plotdata,plotdata$main==""Vitamins and Cofactors"")
+kruskal.test(list(plotdata_vita$SNF1_PT,
+                  plotdata_vita$SNF2_PT,
+                  plotdata_vita$SNF3_PT,
+                  plotdata_vita$SNF4_PT))#p-value = 0.8432
+plotdata_xeno <- subset(plotdata,plotdata$main==""Xenobiotics"")
+kruskal.test(list(plotdata_xeno$SNF1_PT,
+                  plotdata_xeno$SNF2_PT,
+                  plotdata_xeno$SNF3_PT,
+                  plotdata_xeno$SNF4_PT))#p-value = 0.7095
+
+
+plotdata <- gather(plotdata,SNF,abundance,-main,-peak)
+
+library(ggplot2)
+
+#please select optimal color and xlim, ylim
+
+ggplot(plotdata,aes(x = main, y = abundance, fill = SNF)) +
+  scale_fill_manual(values=c(SNF1_PT=""#3D76AE"",SNF2_PT=""#53AD4A"",SNF3_PT=""#EDAB3C"",SNF4_PT=""#C3392A"") )+
+  geom_boxplot(outlier.shape = NA,linetype=""dashed"")+
+  stat_boxplot(outlier.shape=NA,aes(ymin=..lower..,ymax=..upper..))+
+  stat_boxplot(geom = ""errorbar"",aes(ymin=..ymax..),width=0.75)+
+  stat_boxplot(geom = ""errorbar"",aes(ymax=..ymin..),width=0.75)+ 
+  labs(x=""lipid abundance"",y=""lipid abundance"")+
+  # ggtitle(""Boxplot of hub gene"") +
+  theme(panel.grid.major=element_blank(), 
+        panel.grid.minor=element_blank(),
+        panel.background = element_blank())+ 
+  theme(panel.border = element_rect(fill=NA,color=""black"", size=0.8, linetype=""solid"")) 
+
+############## gene heatmap###############
+#""node_anno_rna"":SNF1~3——SNF>0.6; SNF4——SNF<-0.8
+rnalist <- read.csv(""figs2_rnalist.csv"")
+
+rnaheat <- luminal_rna_SNF[rnalist$genelist,1:4]
+
+library(wesanderson)
+n <- rnaheat
+ac=data.frame(group=c(""SNF1"",""SNF2"",""SNF3"",""SNF4""))
+rownames(ac)=colnames(n) 
+P <- pheatmap(as.matrix(n),show_colnames =F,
+              show_rownames = T,
+              scale = ""row"",
+              cluster_rows = F, cluster_cols = F,
+              # main = ""MGY vs MGN"",
+              annotation_col=ac,
+              annotation_names_col =F,
+              annotation_colors = list(group=c(SNF1=""#3D76AE"",SNF2=""#53AD4A"",SNF3=""#EDAB3C"",SNF4=""#C3392A"")),
+              color = rev(colorRampPalette(brewer.pal(9, ""RdBu""))(100))
+              # legend = T,legend_labels=c(""Activity""),
+              # gaps_row=c(9,12)                  
+) 
+P
+
+
+SNF_gene <- as.character(row.names(exp.fpkm.TT))
+KEGG_gene <- as.character(unique(KEGG_Meta_signature$Genes))
+othergene <- setdiff(SNF_gene, KEGG_gene)
+inter_gene <- intersect(SNF_gene, KEGG_gene)
+
+exp.fpkm.TT_kegg <- exp.fpkm.TT[inter_gene,]
+exp.fpkm.TT_normal <- exp.fpkm.PT[inter_gene,]
+
+exp.fpkm.TT_kegg <- log2(cbind(exp.fpkm.TT_kegg,exp.fpkm.TT_normal)+1)
+
+
+# test <- (exp.fpkm.TT_normal[2,1]-exp.fpkm.TT_normal[2,2])^2
+
+
+## Calculate the Euclidean distance between Tumor and Normal
+RMSD <- rep(0, 3861)
+sum <- 0
+x <- 0
+
+for (i in 1:351)
+{
+  for (j in 352:362)
+  {
+    x <- x + 1
+    for (k in 1:nrow(exp.fpkm.TT_kegg))
+    {
+      sum <- sum + (exp.fpkm.TT_kegg[k,i]-exp.fpkm.TT_kegg[k,j])^2
+    }
+    RMSD[x] <- sqrt(sum/nrow(exp.fpkm.TT_kegg))
+    sum <- 0
+  }
+}
+
+write.csv(RMSD,""Metabolic_gene_RMSD_Tumor_Normal.csv"")
+
+## Calculate the Euclidean distance between Normal and Normal
+RMSD <- rep(0, 55)
+sum <- 0
+x <- 0
+
+for (i in 352:361)
+{
+  for (j in (i+1):362)
+  {
+    x <- x + 1
+    for (k in 1:nrow(exp.fpkm.TT_kegg))
+    {
+      sum <- sum + (exp.fpkm.TT_kegg[k,i]-exp.fpkm.TT_kegg[k,j])^2
+    }
+    RMSD[x] <- sqrt(sum/nrow(exp.fpkm.TT_kegg))
+    sum <- 0
+  }
+}
+
+write.csv(RMSD,""Metabolic_gene_RMSD_Normal_Normal.csv"")
+
+## Calculate the Euclidean distance between Tumor and Tumor
+RMSD <- rep(0, 61425)
+sum <- 0
+x <- 0
+
+for (i in 1:350)
+{
+  for (j in (i+1):351)
+  {
+    x <- x + 1
+    for (k in 1:nrow(exp.fpkm.TT_kegg))
+    {
+      sum <- sum + (exp.fpkm.TT_kegg[k,i]-exp.fpkm.TT_kegg[k,j])^2
+    }
+    RMSD[x] <- sqrt(sum/nrow(exp.fpkm.TT_kegg))
+    sum <- 0
+  }
+}
+
+write.csv(RMSD,""Metabolic_gene_RMSD_Tumor_Tumor.csv"")
+
+
+#################plot########################
+
+group <- c(""T VS N"",""T VS T"",""N VS N"")
+
+TVST <- read.csv(""Metabolic_gene_RMSD_Tumor_Tumor.csv"")
+TVST <- as.data.frame(TVST[,2])
+TVST$group <- ""T VS T""
+colnames(TVST) <- c(""Euclidean expression distances"",""group"")
+
+TVSN <- read.csv(""Metabolic_gene_RMSD_Tumor_Normal.csv"")
+TVSN <- as.data.frame(TVSN[,2])
+TVSN$group <- ""T VS N""
+colnames(TVSN) <- c(""Euclidean expression distances"",""group"")
+
+NVSN <- read.csv(""Metabolic_gene_RMSD_Normal_Normal.csv"")
+NVSN <- as.data.frame(NVSN[,2])
+NVSN$group <- ""N VS N""
+colnames(NVSN) <- c(""Euclidean expression distances"",""group"")
+
+plotdata <- rbind(TVST,TVSN,NVSN)
+plotdata[,1] <- as.numeric(plotdata[,1])
+plotdata[,2] <- factor(plotdata[,2],levels = c(""T VS N"",""T VS T"",""N VS N""))
+
+########wilcox##########
+wilcox.test(TVST$`Euclidean expression distances`,TVSN$`Euclidean expression distances`)
+wilcox.test(TVST$`Euclidean expression distances`,NVSN$`Euclidean expression distances`)
+wilcox.test(TVSN$`Euclidean expression distances`,NVSN$`Euclidean expression distances`)
+
+mean(TVST$`Euclidean expression distances`)
+mean(TVSN$`Euclidean expression distances`)
+mean(NVSN$`Euclidean expression distances`)
+
+
+library(ggplot2)
+library(RColorBrewer)
+ggplot(plotdata,aes(x = group, y =`Euclidean expression distances`, fill = group)) +
+  geom_boxplot(width=0.2,outlier.shape = NA,linetype=""dashed"")+
+  stat_boxplot(outlier.shape=NA,aes(ymin=..lower..,ymax=..upper..))+
+  stat_boxplot(geom = ""errorbar"",aes(ymin=..ymax..),width=0.3)+
+  stat_boxplot(geom = ""errorbar"",aes(ymax=..ymin..),width=0.3)+ 
+  scale_fill_manual(values = colorRampPalette(brewer.pal(6, ""Spectral""))(3))+
+  scale_y_continuous(expand = c(0,0),limits = c(0,1.6))+
+  labs(x=""SNF subtype"",y=""Distribution distance (r.m.s.d.)"")+
+  theme(panel.grid.major=element_blank(), 
+        panel.grid.minor=element_blank(),
+        panel.background = element_blank())+
+  theme(panel.border = element_rect(fill=NA,color=""black"", size=0.8, linetype=""solid"")) 
+
+
+##############################################################################
+{
+  ## Calculate the Euclidean distance of all genes
+  RMSD <- rep(0, 8680)
+  sum <- 0
+  x <- 0
+  
+  for (i in 1:360)
+  {
+    for (j in 361:448)
+    {
+      x <- x + 1
+      for (k in 1:nrow(FUSCCTNBC_FPKM_log_trans))
+      {
+        sum <- sum + (FUSCCTNBC_FPKM_log_trans[k,i]-FUSCCTNBC_FPKM_log_trans[k,j])^2
+      }
+      RMSD[x] <- sqrt(sum/nrow(FUSCCTNBC_FPKM_log_trans))
+      sum <- 0
+    }
+  }
+  
+  write.csv(RMSD,""All_gene_RMSD_Tumor_Normal.csv"")
+  
+  RMSD <- rep(0, 3828)
+  sum <- 0
+  x <- 0
+  
+  for (i in 361:447)
+  {
+    for (j in (i+1):448)
+    {
+      x <- x + 1
+      for (k in 1:nrow(FUSCCTNBC_FPKM_log_trans))
+      {
+        sum <- sum + (FUSCCTNBC_FPKM_log_trans[k,i]-FUSCCTNBC_FPKM_log_trans[k,j])^2
+      }
+      RMSD[x] <- sqrt(sum/nrow(FUSCCTNBC_FPKM_log_trans))
+      sum <- 0
+    }
+  }
+  
+  write.csv(RMSD,""All_gene_RMSD_Normal_Normal.csv"")
+  
+  RMSD <- rep(0, 64620)
+  sum <- 0
+  x <- 0
+  
+  for (i in 1:359)
+  {
+    for (j in (i+1):360)
+    {
+      x <- x + 1
+      for (k in 1:nrow(FUSCCTNBC_FPKM_log_trans))
+      {
+        sum <- sum + (FUSCCTNBC_FPKM_log_trans[k,i]-FUSCCTNBC_FPKM_log_trans[k,j])^2
+      }
+      RMSD[x] <- sqrt(sum/nrow(FUSCCTNBC_FPKM_log_trans))
+      sum <- 0
+    }
+  }
+  
+  write.csv(RMSD,""All_gene_RMSD_Tumor_Tumor.csv"")
+}
+
+##############SNF##################
+# Calculate Euclidean distance for BCMA expression data
+exp.fpkm.TT_kegg
+
+BCMA_RNA_Combat_log2_FPKM_PT <- exp.fpkm.TT_kegg[,352:362] #11
+BCMA_RNA_Combat_log2_FPKM_SNF1 <- exp.fpkm.TT_kegg[,intersect(colnames(exp.fpkm.TT_kegg),SNF1_ID)] # 86
+BCMA_RNA_Combat_log2_FPKM_SNF2 <- exp.fpkm.TT_kegg[,intersect(colnames(exp.fpkm.TT_kegg),SNF2_ID)] # 89
+BCMA_RNA_Combat_log2_FPKM_SNF3 <- exp.fpkm.TT_kegg[,intersect(colnames(exp.fpkm.TT_kegg),SNF3_ID)] #118
+BCMA_RNA_Combat_log2_FPKM_SNF4 <- exp.fpkm.TT_kegg[,intersect(colnames(exp.fpkm.TT_kegg),SNF4_ID)] #58
+
+## Calculate the Euclidean distance between Tumor and Normal
+
+comparison_Lum_SNF2E_SNF1_distance <- matrix(ncol=4,nrow=nrow(BCMA_RNA_Combat_log2_FPKM_SNF1))
+colnames(comparison_Lum_SNF2E_SNF1_distance) <- c(""SNF3"",""SNF4"",""SNF2"",""SNF1"")
+rownames(comparison_Lum_SNF2E_SNF1_distance) <- rownames(BCMA_RNA_Combat_log2_FPKM_SNF1)
+
+## SNF1 vs PT
+RMSD <- rep(0, ncol(BCMA_RNA_Combat_log2_FPKM_SNF1)*11)
+sum <- 0
+x <- 0
+
+for (i in 1:ncol(BCMA_RNA_Combat_log2_FPKM_SNF1)){
+  for (j in 1:11){
+    x <- x + 1
+    for (k in 1:nrow(comparison_Lum_SNF2E_SNF1_distance)){
+      sum <- sum + (BCMA_RNA_Combat_log2_FPKM_SNF1[k,i]-BCMA_RNA_Combat_log2_FPKM_PT[k,j])^2}
+    RMSD[x] <- sqrt(sum/nrow(comparison_Lum_SNF2E_SNF1_distance))
+    sum <- 0}
+  print(i)}
+RMSD_SNF1_PT <- c(RMSD)
+length(RMSD_SNF1_PT) <- 946
+
+
+## SNF2 vs PT
+RMSD <- rep(0, ncol(BCMA_RNA_Combat_log2_FPKM_SNF2)*11)
+sum <- 0
+x <- 0
+
+for (i in 1:ncol(BCMA_RNA_Combat_log2_FPKM_SNF2)){
+  for (j in 1:11){
+    x <- x + 1
+    for (k in 1:nrow(comparison_Lum_SNF2E_SNF1_distance)){
+      sum <- sum + (BCMA_RNA_Combat_log2_FPKM_SNF2[k,i]-BCMA_RNA_Combat_log2_FPKM_PT[k,j])^2}
+    RMSD[x] <- sqrt(sum/nrow(comparison_Lum_SNF2E_SNF1_distance))
+    sum <- 0}
+  print(i)}
+RMSD_SNF2_PT <- c(RMSD)
+length(RMSD_SNF2_PT) <- 979
+
+
+## SNF3 vs PT
+RMSD <- rep(0, ncol(BCMA_RNA_Combat_log2_FPKM_SNF3)*11)
+sum <- 0
+x <- 0
+
+for (i in 1:ncol(BCMA_RNA_Combat_log2_FPKM_SNF3)){
+  for (j in 1:11){
+    x <- x + 1
+    for (k in 1:nrow(comparison_Lum_SNF2E_SNF1_distance)){
+      sum <- sum + (BCMA_RNA_Combat_log2_FPKM_SNF3[k,i]-BCMA_RNA_Combat_log2_FPKM_PT[k,j])^2}
+    RMSD[x] <- sqrt(sum/nrow(comparison_Lum_SNF2E_SNF1_distance))
+    sum <- 0}
+  print(i)}
+RMSD_SNF3_PT <- c(RMSD)
+length(RMSD_SNF3_PT) <- 1298
+
+
+## SNF4 vs PT
+RMSD <- rep(0, ncol(BCMA_RNA_Combat_log2_FPKM_SNF4)*11)
+sum <- 0
+x <- 0
+
+for (i in 1:ncol(BCMA_RNA_Combat_log2_FPKM_SNF4)){
+  for (j in 1:11){
+    x <- x + 1
+    for (k in 1:nrow(comparison_Lum_SNF2E_SNF1_distance)){
+      sum <- sum + (BCMA_RNA_Combat_log2_FPKM_SNF4[k,i]-BCMA_RNA_Combat_log2_FPKM_PT[k,j])^2}
+    RMSD[x] <- sqrt(sum/nrow(comparison_Lum_SNF2E_SNF1_distance))
+    sum <- 0}
+  print(i)}
+RMSD_SNF4_PT <- c(RMSD)
+length(RMSD_SNF4_PT) <- 638
+
+
+#### GGPLOT2
+SNF3_PT_all <- matrix(ncol=2,nrow=length(RMSD_SNF3_PT))
+SNF3_PT_all[,1] <- ""SNF3""
+SNF3_PT_all[,2] <- RMSD_SNF3_PT
+
+SNF4_PT_all <- matrix(ncol=2,nrow=length(RMSD_SNF4_PT))
+SNF4_PT_all[,1] <- ""SNF4""
+SNF4_PT_all[,2] <- RMSD_SNF4_PT
+
+SNF1_PT_all <- matrix(ncol=2,nrow=length(RMSD_SNF1_PT))
+SNF1_PT_all[,1] <- ""SNF1""
+SNF1_PT_all[,2] <- RMSD_SNF1_PT
+
+SNF2_PT_all <- matrix(ncol=2,nrow=length(RMSD_SNF2_PT))
+SNF2_PT_all[,1] <- ""SNF2""
+SNF2_PT_all[,2] <- RMSD_SNF2_PT
+
+comparison_SNF_BCMA_ggboxplot <- rbind(SNF3_PT_all,SNF4_PT_all,SNF2_PT_all,SNF1_PT_all)
+colnames(comparison_SNF_BCMA_ggboxplot) <- c(""SNF"",""RMSD"")
+comparison_SNF_BCMA_ggboxplot <- as.data.frame(comparison_SNF_BCMA_ggboxplot)
+comparison_SNF_BCMA_ggboxplot[,2] <- as.numeric(comparison_SNF_BCMA_ggboxplot[,2])
+
+#
+comparison_SNF_BCMA_ggboxplot$SNF <- factor(comparison_SNF_BCMA_ggboxplot$SNF,levels = c(""SNF1"",""SNF2"",""SNF3"",""SNF4""))
+
+ggplot(comparison_SNF_BCMA_ggboxplot,
+       aes(x=SNF,y=RMSD,fill=SNF))+
+  #scale_x_discrete(limits=c(""polar"",""lipid"",""merge""))+
+  scale_fill_manual(values=c(SNF1=""#3D76AE"",SNF2=""#53AD4A"",SNF3=""#EDAB3C"",SNF4=""#C3392A"") )+
+  geom_boxplot(outlier.shape = NA,linetype=""dashed"")+
+  stat_boxplot(outlier.shape=NA,aes(ymin=..lower..,ymax=..upper..))+
+  stat_boxplot(geom = ""errorbar"",aes(ymin=..ymax..),width=0.3)+
+  stat_boxplot(geom = ""errorbar"",aes(ymax=..ymin..),width=0.3)+ 
+  scale_y_continuous(expand = c(0,0.1),limits = c(0.3,1.6))+
+  labs(x=""SNF subtype"",y=""Distribution distance (r.m.s.d.)"")+
+  theme(panel.grid.major=element_blank(), 
+        panel.grid.minor=element_blank(),
+        panel.background = element_blank())+ 
+  theme(panel.border = element_rect(fill=NA,color=""black"", size=0.8, linetype=""solid""))
+#stat_compare_means(hide.ns = TRUE,label = ""p.signif"",method=""wilcox.test"")
+#geom_signif(comparisons = compaired,step_increase = 0.06,map_signif_level = T,test = wilcox.test) 
+
+mean(as.numeric(SNF1_PT_all[,2]))#0.7590856
+mean(as.numeric(SNF2_PT_all[,2]))#0.7675502
+mean(as.numeric(SNF3_PT_all[,2]))#0.8377549
+mean(as.numeric(SNF4_PT_all[,2]))#0.6759793
+#1/2=99%
+#1/3=91%
+#1/4=89%
+#2/3=92%
+#2/4=88%
+#3/4=81%
+
+wilcox.test(as.numeric(SNF1_PT_all[,2]),as.numeric(SNF2_PT_all[,2]))
+#p-value = 0.00412
+wilcox.test(as.numeric(SNF1_PT_all[,2]),as.numeric(SNF3_PT_all[,2]))
+#p-value < 2.2e-16
+wilcox.test(as.numeric(SNF1_PT_all[,2]),as.numeric(SNF4_PT_all[,2]))
+#p-value < 2.2e-16
+wilcox.test(as.numeric(SNF2_PT_all[,2]),as.numeric(SNF3_PT_all[,2]))
+#p-value < 2.2e-16
+wilcox.test(as.numeric(SNF2_PT_all[,2]),as.numeric(SNF4_PT_all[,2]))
+#p-value < 2.2e-16
+wilcox.test(as.numeric(SNF3_PT_all[,2]),as.numeric(SNF4_PT_all[,2]))
+#p-value < 2.2e-16
+
+kruskal.test(list(as.numeric(SNF1_PT_all[,2]),
+                  as.numeric(SNF2_PT_all[,2]),
+                  as.numeric(SNF3_PT_all[,2]),
+                  as.numeric(SNF4_PT_all[,2])))
+#p-value < 2.2e-16
+
+{
+  ggplot(plotdata,aes(x = group, y =plotdata$`Euclidean expression distances`, fill = group)) +
+    stat_boxplot(geom = ""errorbar"", width=0.6)+
+    geom_boxplot(width = 0.6, outlier.shape=NA)+
+    scale_fill_manual(values = colorRampPalette(brewer.pal(6, ""Spectral""))(3))+
+    scale_y_continuous(expand = c(0,0),limits = c(0,2))+
+    # scale_x_discrete(name = ""lipid category"") +
+    # ggtitle(""Boxplot of hub gene"") +
+    theme_bw() +
+    theme(plot.title = element_text(size = 14, face =  ""bold""),
+          text = element_text(size = 12),
+          axis.title = element_text(face=""bold""),
+          axis.text.x=element_text(size = 11),
+          panel.grid.major=element_line(colour=NA),
+          panel.background = element_rect(fill = ""transparent"",colour = NA),
+          plot.background = element_rect(fill = ""transparent"",colour = NA),
+          panel.grid.minor = element_blank())
+}
+
+mut_met <- read.csv(""mut_met.csv"")
+pol_anno <- CBCGA_pol_anno[,c(1,5,6)]
+mut_met <- merge(mut_met,pol_anno,by.x = ""metabolite"",by.y = ""peak"",all = F)
+write.csv(mut_met,""mut_met_anno.csv"")
+
+mut_lip <- read.csv(""mut_lip.csv"")
+pol_anno <- CBCGA_lip_anno[,c(2,4)]
+mut_lip <- merge(mut_lip,pol_anno,by.x = ""metabolite"",by.y = ""Subclass"",all = F)
+write.csv(mut_lip,""mut_lip_anno.csv"")
+
+rna_met <- read.csv(""rna_met.csv"")
+pol_anno <- CBCGA_pol_anno[,c(1,5,6)]
+rna_met <- merge(rna_met,pol_anno,by.x = ""metabolite"",by.y = ""peak"",all = F)
+write.csv(rna_met,""rna_met_anno.csv"")
+
+cnv_met <- read.csv(""cnv_met.csv"")
+pol_anno <- CBCGA_pol_anno[,c(1,5,6)]
+cnv_met <- merge(cnv_met,pol_anno,by.x = ""metabolite"",by.y = ""peak"",all = F)
+write.csv(cnv_met,""cnv_met_anno.csv"")
+
+cyc_met <- read.csv(""cyc_met.csv"")
+pol_anno <- CBCGA_pol_anno[,c(1,5,6)]
+cyc_met <- merge(cyc_met,pol_anno,by.x = ""metabolite"",by.y = ""peak"",all = F)
+write.csv(cyc_met,""cyc_met_anno.csv"")
+
+luminal_Pol_SNF$metabolite <- rownames(luminal_Pol_SNF)
+luminal_Pol_SNF <- merge(luminal_Pol_SNF,pol_anno,by.x = ""metabolite"",by.y = ""peak"",all = F)
+write.csv(luminal_Pol_SNF,""luminal_Pol_SNF.csv"")
+
+#########################Plot############################################
+
+library(ggplot2)
+library(ggpubr)
+library(gridExtra)
+
+p1 <- ggplot(Customed_mutation_plus_metabolomics, aes(x=order,y=M613T483_POS, colour = mutation, shape = mutation))+
+  geom_point(size = 3) + theme_bw() + ylab(""Log2 levels"") + scale_color_manual(values = c(""black"",""red"")) + scale_shape_manual(values = c(16,17)) +
+  theme(panel.grid =element_blank(), axis.text.x = element_blank(), axis.ticks.x = element_blank(), legend.position=""none"") 
+
+p2 <- ggplot(Customed_mutation_plus_metabolomics, aes(x = order, y = 0)) + 
+  geom_tile(aes(fill = mutation)) + theme_bw() + xlab(""Ordered samples"") + ylab(""mutation mutations"") +
+  theme(panel.grid =element_blank(), axis.text = element_blank(), axis.ticks = element_blank(), legend.position=""none"") +
+  scale_fill_manual(values = c(""white"",""black""))
+
+gp1<- ggplot_gtable(ggplot_build(p1))
+gp2<- ggplot_gtable(ggplot_build(p2))
+#This identifies the maximum width
+maxWidth = unit.pmin(gp1$widths[2:3], gp2$widths[2:3])
+#Set each to the maximum width
+gp1$widths[2:3] <- maxWidth
+gp2$widths[2:3] <- maxWidth
+#Put them together
+grid.arrange(gp1, gp2)
+
+
+
+
+","R"
+"Receptor","yifanzhou330/Luminal-SNF","Fig2_and_related_ED_Fig/Oncoplot.R",".R","9670","238","rm(list = ls())
+graphics.off()
+library(maftools)
+library(ComplexHeatmap)
+library(circlize)
+library(RColorBrewer)
+
+load(""./CBCGA_HRposHER2neg351_WES_RNAseq_CNV_Metab_Protein_20220829.Rdata"")
+luminal$SNF = paste0(""SNF"",SNF_Cluster[rownames(luminal)])
+
+######## annot prepare -------
+luminal$HRD = as.numeric(luminal$HRD)
+
+######## prepre mut plot  --------
+luminal_mut = luminal
+luminal_mut$Tumor_Sample_Barcode = luminal_mut$PatientCode
+luminal_mut = luminal_mut[intersect(rownames(luminal_mut),unique(maf.df$Tumor_Sample_Barcode)),]
+SNF_Cluster_mut = SNF_Cluster[rownames(luminal_mut)]
+
+maf.df = maf.df[maf.df$Tumor_Sample_Barcode %in% rownames(luminal_mut),]
+SNF_maf = read.maf(maf.df,clinicalData = luminal_mut)
+
+oncoplot(SNF_maf,top = 10,
+         clinicalFeatures = ""SNF"",sortByAnnotation = T,writeMatrix = T,removeNonMutated = F,
+         groupAnnotationBySize = F)
+mut_mtx0 = read.table(""./onco_matrix.txt"",sep = ""\t"", na.strings = c("""",0))
+tmp = matrix(""Unknown"",nrow = nrow(mut_mtx0), ncol = nrow(luminal) -ncol(mut_mtx0) )
+rownames(tmp) = rownames(mut_mtx0) ; colnames(tmp) = setdiff(rownames(luminal), colnames(mut_mtx0))
+mut_mtx0 = cbind(mut_mtx0,tmp)
+genelist1 = read.csv(""/Census_allMon Jan 11 04_06_52 2021.csv"") 
+genelist1 = unique(genelist1$Gene.Symbol)
+genelist2 = data.table::fread(""NCG6_cancergenes.tsv"",data.table = F) 
+genelist2 = unique(genelist2$symbol)
+genelist3 = data.table::fread(""./FLAGS supplementary.txt"",data.table = F)
+genelist3 = genelist3$FLAGS
+genelist = setdiff(intersect(genelist1, genelist2),genelist3)
+genelist = intersect(rownames(mut_mtx0),genelist)
+mut_mtx0 = mut_mtx0[genelist,rownames(luminal)]
+
+mut_mtx =  mutCountMatrix(SNF_maf, includeSyn = F, countOnly = NULL, removeNonMutated = F)
+mut_mtx = mut_mtx_stat = apply(mut_mtx, 2, function(x){ifelse(x == 0, ""WT"", ""Mut"")} )
+tmp = matrix(""Unknown"",nrow = nrow(mut_mtx), ncol = nrow(luminal) -ncol(mut_mtx) )
+rownames(tmp) = rownames(mut_mtx) ; colnames(tmp) = setdiff(rownames(luminal), colnames(mut_mtx))
+mut_mtx = cbind(mut_mtx,tmp)
+mut_mtx = mut_mtx[ genelist ,rownames(luminal)]
+mut_mtx_stat = mut_mtx_stat[genelist,]
+
+######## define order -------
+tmp = luminal ; tmp$Tumor_Sample_Barcode = tmp$PatientCode
+tmp = tmp[unique(maf.df$Tumor_Sample_Barcode),]
+tmp = read.maf(maf.df,clinicalData = tmp)
+oncoplot(tmp,genes = rownames(mut_mtx),clinicalFeatures = ""SNF"",sortByAnnotation = T,writeMatrix = T,removeNonMutated = F,
+         groupAnnotationBySize = F)
+tmp = read.table(""./onco_matrix.txt"",sep = ""\t"")
+order = colnames(tmp)
+no_mut = luminal[setdiff(rownames(luminal),order),""SNF""]
+names(no_mut) = setdiff(rownames(luminal),order)
+
+tmp = luminal[order,""SNF""]
+names(tmp) = order
+
+order = c(tmp,no_mut)
+order = sort(order)
+
+
+######## prepare cnv plot  --------
+amp_mtx = cnv.alldata[c(""CCND1"",""MDM2"",""FGFR1""),]
+amp_mtx = apply(amp_mtx, 2, function(x){ifelse(x < log2(4/2), ""non_Amp"", ""Amp"")} )
+amp_mtx = amp_mtx[,names(order)]
+
+
+######### merge & draw ------
+color_PAM50 = c(""#1D76BC"",""#76CFE6"",""#6E59A6"",""#E11D2E"",""#CDCFD0"")
+names(color_PAM50) = c(""LumA"",""LumB"",""Her2"",""Basal"",""Normal"")
+color_SNF = color
+names(color_SNF) = paste0(""SNF"",1:4)
+
+col_fun_Isig <- colorRamp2(c(-0.5, 0.3 ,2), colorRampPalette(c(""white"", ""#08506F""))(20)[c(2,10,20)])
+col_fun_SSig <- colorRamp2(c(-0.1, 0.75 ,1.2), colorRampPalette(c(""white"", ""#50A9AC""))(20)[c(2,10,20)])
+col_fun_hrd <- colorRamp2(c(0, 16, 75), c('#FDF9E7', '#EFE089', '#D7C801'))
+
+#### mut ------
+col = c(""Missense_Mutation"" = ""#366A9C"", ""Nonsense_Mutation"" = ""#BBDE93"",
+        ""In_Frame_Del"" = ""#EE8632"" ,""In_Frame_Ins"" = ""#D0342B"",
+        ""Multi_Hit"" = ""black"", ""Splice_Site"" = ""#F3C17B"", ""Frame_Shift_Del"" = ""#AECDE1"",
+        ""Frame_Shift_Ins"" = ""#ED9E9B"",
+        ""Nonstop_Mutation"" = ""#339900"",
+        ""Unknown"" = ""#eaeaea"")
+wid = 3
+hei = 2
+
+alter_fun <- list(
+  background = function(x, y, w, h) {
+    grid.rect(x, y, w-unit(wid, ""pt""), h-unit(hei, ""pt""),
+              gp = gpar(fill = ""white"", col = ""white""))
+  },
+  Missense_Mutation = function(x, y, w, h) {
+    grid.rect(x, y, w-unit(wid, ""pt""), h-unit(hei, ""pt""),
+              gp = gpar(fill = col[""Missense_Mutation""], col = col[""Missense_Mutation""]))
+  },
+  Nonsense_Mutation = function(x, y, w, h) {
+    grid.rect(x, y, w-unit(wid, ""pt""), h-unit(hei, ""pt""),
+              gp = gpar(fill = col[""Nonsense_Mutation""], col = col[""Nonsense_Mutation""]))
+  },
+  In_Frame_Del = function(x, y, w, h) {
+    grid.rect(x, y, w-unit(wid, ""pt""), h-unit(hei, ""pt""),
+              gp = gpar(fill = col[""In_Frame_Del""], col = col[""In_Frame_Del""]))
+  },
+  In_Frame_Ins = function(x, y, w, h) {
+    grid.rect(x, y, w-unit(wid, ""pt""), h-unit(hei, ""pt""),
+              gp = gpar(fill = col[""In_Frame_Ins""], col = col[""In_Frame_Ins""]))
+  },
+  Multi_Hit = function(x, y, w, h) {
+    grid.rect(x, y, w-unit(wid, ""pt""), h-unit(hei, ""pt""),
+              gp = gpar(fill = col[""Multi_Hit""], col = col[""Multi_Hit""]))
+  },
+  Splice_Site = function(x, y, w, h) {
+    grid.rect(x, y, w-unit(wid, ""pt""), h-unit(hei, ""pt""),
+              gp = gpar(fill = col[""Splice_Site""], col = col[""Splice_Site""]))
+  },
+  Frame_Shift_Del = function(x, y, w, h) {
+    grid.rect(x, y, w-unit(wid, ""pt""), h-unit(hei, ""pt""),
+              gp = gpar(fill = col[""Frame_Shift_Del""], col = col[""Frame_Shift_Del""]))
+  },
+  Frame_Shift_Ins = function(x, y, w, h) {
+    grid.rect(x, y, w-unit(wid, ""pt""), h-unit(hei, ""pt""),
+              gp = gpar(fill = col[""Frame_Shift_Ins""], col = col[""Frame_Shift_Ins""]))
+  },
+  Nonstop_Mutation = function(x, y, w, h) {
+    grid.rect(x, y, w-unit(wid, ""pt""), h-unit(hei, ""pt""),
+              gp = gpar(fill = col[""Nonstop_Mutation""], col = col[""Nonstop_Mutation""]))
+  },
+  Unknown = function(x, y, w, h) {
+    grid.rect(x, y, w-unit(wid, ""pt""), h-unit(hei, ""pt""),
+              gp = gpar(fill = col[""Unknown""], col = col[""Unknown""]))
+  }
+)
+
+luminal = luminal[names(order),]
+p1 <- oncoPrint(mut_mtx0[,names(order)],  
+                alter_fun = alter_fun, col = col, show_pct = T, pct_side = 'right',
+                column_title = '', column_order = names(order),
+                row_names_side = ""left"", show_column_names = F, remove_empty_columns = F,
+                heatmap_legend_param = list(title = ""Mut""),border = T,
+                row_names_gp = gpar(fontsize = 9), column_title_gp = gpar(fontsize = 0),
+                column_split = factor(luminal[names(order),""SNF""], levels = c('SNF1', 'SNF2', 'SNF3', 'SNF4')),column_gap = unit(2, ""mm""),
+                top_annotation = HeatmapAnnotation(SNF = luminal$SNF,
+                                                   HRD = luminal$HRD,
+                                                   annotation_name_side = ""left"", annotation_name_gp = gpar(fontsize = 9),
+                                                   col = list( SNF = color_SNF, HRD = col_fun_hrd) ) )
+
+#### cnv ------
+amp_mtx2 = amp_mtx
+for (i in colnames(amp_mtx2)){
+  mat = amp_mtx2[,i]
+  mat[mat == ""non_Amp""] = """"
+  amp_mtx2[,i] = mat
+}
+
+col2 = c( ""Amp"" = ""#BC102B"")
+wid = 3
+hei = 2
+alter_fun2 <- list(
+  background = function(x, y, w, h) {
+    grid.rect(x, y, w-unit(wid, ""pt""), h-unit(hei, ""pt""),
+              gp = gpar(fill = ""white"", col = ""white""))
+  },
+  Amp = function(x, y, w, h) {
+    grid.rect(x, y, w-unit(wid, ""pt""), h-unit(hei, ""pt""),
+              gp = gpar(fill = col2[""Amp""], col = col2[""Amp""]))
+  }
+  )
+p2 = oncoPrint(amp_mtx2[,names(order)],  
+               alter_fun = alter_fun2,col = col2,
+             show_pct = T, 
+             show_column_names = FALSE,
+             na_col = ""#eaeaea"", border = T, 
+             row_names_gp = gpar(fontsize = 9),
+             heatmap_legend_param = list(title = ""CNV""),
+             row_names_side = ""left"",
+             pct_side = 'right',
+             top_annotation = NULL,
+             column_split = factor(luminal[names(order),""SNF""], levels = c('SNF1', 'SNF2', 'SNF3', 'SNF4')),
+             column_gap = unit(2, ""mm"")
+)
+
+hp_list = p1 %v% p2
+export::graph2pdf(hp_list,file =  ""Mut_CNA_plot.pdf"",width = 14, height = 7)
+
+
+######### stat ------------
+filename = ""Mut_CNA_plot""
+
+## mut
+stat_res = as.data.frame(matrix(nrow = nrow(mut_mtx_stat),ncol = 6))
+rownames(stat_res) = rownames(mut_mtx_stat)
+colnames(stat_res) = c(""SNF1"",""SNF2"",""SNF3"",""SNF4"",""p.val"",""adj.p"")
+for( i in rownames(stat_res)){
+  mat = data.frame(gene = mut_mtx_stat[i,names(sort(SNF_Cluster_mut))], 
+                   group = paste0(""SNF"",sort(SNF_Cluster_mut)))
+  res = chisq.test(table(mat$gene,mat$group))
+  tmp = as.matrix.data.frame(table(mat$gene,mat$group))
+  
+  stat_res[i,""SNF1""] = tmp[1,1]/sum(tmp[,1])
+  stat_res[i,""SNF2""] = tmp[1,2]/sum(tmp[,2])
+  stat_res[i,""SNF3""] = tmp[1,3]/sum(tmp[,3])
+  stat_res[i,""SNF4""] = tmp[1,4]/sum(tmp[,4])
+  stat_res[i,""p.val""] = res[[""p.value""]]
+}
+stat_res[,""adj.p""] = p.adjust(stat_res[,""p.val""],method = ""fdr"")
+write.csv(stat_res, file =  paste0(filename,""_mut_chisq.csv""))
+
+## CNA
+stat_res = as.data.frame(matrix(nrow = nrow(amp_mtx),ncol = 6))
+rownames(stat_res) = rownames(amp_mtx)
+colnames(stat_res) = c(""SNF1"",""SNF2"",""SNF3"",""SNF4"",""p.val"",""adj.p"")
+for( i in rownames(stat_res)){
+  mat = data.frame(gene = amp_mtx[i,names(sort(SNF_Cluster))], 
+                   group = paste0(""SNF"",sort(SNF_Cluster)))
+  res = chisq.test(table(mat$gene,mat$group))
+  
+  tmp = as.matrix.data.frame(table(mat$gene,mat$group))
+  stat_res[i,""SNF1""] = tmp[1,1]/sum(tmp[,1])
+  stat_res[i,""SNF2""] = tmp[1,2]/sum(tmp[,2])
+  stat_res[i,""SNF3""] = tmp[1,3]/sum(tmp[,3])
+  stat_res[i,""SNF4""] = tmp[1,4]/sum(tmp[,4])
+  stat_res[i,""p.val""] = res[[""p.value""]]
+}
+stat_res[,""adj.p""] = p.adjust(stat_res[,""p.val""],method = ""fdr"")
+write.csv(stat_res, file =  paste0(filename,""_AMP_chisq.csv""))
+
+
+
+
+
+
+
+","R"
+"Receptor","yifanzhou330/Luminal-SNF","Fig3_and_related_ED_Fig/External_Validation.R",".R","3082","87","rm(list = ls())  ; graphics.off()
+library(xlsx)
+library(GSVA)
+library(GSEABase)
+library(RColorBrewer)
+library(data.table)
+library(stringi)  ; library(stringr)
+library(limma)
+library(ggpubr)
+library(plyr)
+library(export)
+library(ComplexHeatmap) ; library(circlize)
+test_summary = openxlsx:::read.xlsx(""./RawNum100_CorThresh0.75_RF_scale_fold0_CPTAC.xlsx"")
+rownames(test_summary) = test_summary$X1
+
+
+clinic = read.csv(""./prosp-brca-v5.4-public-sample-annotation.csv"",
+                  row.names = 1)
+clinic = clinic[rownames(test_summary),]
+clinic$SNF =  test_summary$inferSNF
+clinic$SNF = factor(clinic$SNF, levels = paste0(""SNF"",1:4))
+
+rna = data.table::fread(""./prosp-brca-v5.4-public-rnaseq-fpkm-log2.csv"",
+                        data.table = F)
+rna = rna[!duplicated(rna$V1),]
+rownames(rna) = rna[,1]
+rna = rna[,-1]
+rna = rna[,rownames(test_summary)]
+
+cnv = read.csv(""./prosp-brca-v5.4-public-gene-level-cnv-gistic2-all_data_by_genes.gct.csv"")
+cnv = cnv[!duplicated(cnv$geneSymbol),]
+rownames(cnv) = cnv$geneSymbol
+cnv = cnv[,-c(1:4)]
+cnv = cnv[,rownames(test_summary)]
+
+clinic$Stromal.Score = clinic$xCell.Stromal.Score
+clinic$Immune.Score = clinic$xCell.Immune.Score
+
+clinic$CINscore = NA
+for (i in rownames(clinic)){
+  tmp = cnv[,i]
+  tmp = tmp^2
+  clinic[i,""CINscore""] = sum(tmp)
+}
+clinic$CINscore = log2(clinic$CINscore)
+
+gmt = getGmt(""./GObp_Hall_Reactome_GOmf_v7_4.gmt"") 
+tmp = c(""HALLMARK_ESTROGEN_RESPONSE_EARLY"")#,""HALLMARK_ESTROGEN_RESPONSE_LATE"")
+gmt_picked = gmt[tmp]
+path_sig = gsva(as.matrix(rna),gmt_picked,method = ""gsva"")#,kcdf = ""Gaussian"",abs.ranking=F)
+
+#clinic = cbind(clinic,as.data.frame(t(path_sig[,rownames(clinic)])))
+clinic = cbind(clinic,ESTROGEN_RESPONSE_EARLY = as.numeric(path_sig[,rownames(clinic)]) )
+#########
+clinic = clinic[,c(""SNF"",""Stromal.Score"", ""CINscore"",""Immune.Score"",
+                   ""ESTROGEN_RESPONSE_EARLY"")]
+colnames(clinic)[c(2,4)] = c(""Stromal.Score"", ""Immune.Score"") 
+mat = aggregate(clinic[,-1],list(clinic$SNF),median)
+rownames(mat) = mat$Group.1
+mat = as.data.frame(t(mat[,-1]))
+
+
+col_fun = colorRamp2(c(-0.6,-0.5, 0 , 0.5,0.6), rev(brewer.pal(n = 5, name =""RdBu"")))
+p=ComplexHeatmap::pheatmap(mat,scale = ""row"",cluster_rows = F,cluster_cols = F,
+                           heatmap_legend_param = list(title = ""Z-score""),
+                           border_color = ""white"",
+                           color = col_fun)
+graph2pdf(p,file = ""./CPTAC_molecular_feature.pdf"",width = 5,height = 7)
+
+
+######### kruskal.test
+kruskal.test.res = as.data.frame(matrix(nrow = nrow(mat),ncol = 2))
+rownames(kruskal.test.res) = rownames(mat) ; colnames(kruskal.test.res) = c(""p.val"",'adj.p')
+for ( i in rownames(mat)){
+  tmp = data.frame(gene = as.numeric(clinic[,i]),
+                   group = clinic$SNF)
+  res = kruskal.test(gene~group,data = tmp)
+  kruskal.test.res[i,""p.val""] = res[[""p.value""]]
+  # res = summary(aov(gene~group,data = tmp))
+  # aov.res[i,""p.val""] = res[[1]][[""Pr(>F)""]][1]
+}
+kruskal.test.res$adj.p = p.adjust(kruskal.test.res$p.val,method = ""fdr"")
+
+write.csv(kruskal.test.res,file = ""./kruskal.test.res.csv"")
+
+
+","R"
+"Receptor","yifanzhou330/Luminal-SNF","Fig3_and_related_ED_Fig/Survival_analysis.R",".R","4604","100","rm(list = ls()) ; graphics.off()
+library(ggplot2)
+library(survminer)
+library(survival)
+library(""export"")
+library(forestplot)
+load(""CBCGA_HRposHER2neg351_WES_RNAseq_CNV_Metab_Protein_20220829.Rdata"")
+luminal$DMFS_status = as.numeric(luminal$DMFS_status)
+luminal$DMFS_months = as.numeric(luminal$DMFS_months)
+luminal$RFS_status = as.numeric(luminal$RFS_status)
+luminal$RFS_months = as.numeric(luminal$RFS_months)
+luminal$SNF = paste0(""SNF"", SNF_Cluster[rownames(luminal)])
+luminal$PAM50 = factor(luminal$PAM50_classifier,levels = c(""LumA"",""LumB"",""Her2"",""Basal"",""Normal""))
+color_PAM50 = c(Normal = ""#CDCFD0"" , LumB = ""#76CFE6"" , LumA = ""#1D76BC"" , Her2  = ""#6E59A6"", Basal = ""#E11D2E"")
+
+############################ km plot -----
+for ( i in unique(luminal$PAM50_classifier)){
+  mat = luminal[luminal$PAM50_classifier == i,]
+  km = survfit(Surv(DMFS_months,DMFS_status) ~ SNF,data = mat)
+  p = ggsurvplot(km,pval = T,pval.method = T,palette =  c(""#2378B3"",  ""#1CB038"",  ""#F8A900"", ""#D5271A""),
+             break.x.by = 12)
+  ggsave(p$plot,filename = paste0(""surplot_SNF_DMFS_in_"",i,"".pdf""),height = 6,width = 6)
+  graph2ppt(p$plot,file =paste0(""surplot_SNF_DMFS_in_"",i,"".pptx""),height = 6,width = 6)
+
+}
+
+km = survfit(Surv(DMFS_months,DMFS_status) ~ SNF,data = luminal)
+p = ggsurvplot(km,pval = T,pval.method = T,palette =  c(""#2378B3"",  ""#1CB038"",  ""#F8A900"", ""#D5271A""),
+               break.x.by = 12)
+ggsave(p$plot,filename = paste0(""surplot_SNF_DMFS.pdf""),height = 6,width = 6)
+graph2ppt(p$plot,file =paste0(""surplot_SNF_DMFS.pptx""),height = 6,width = 6)
+
+km = survfit(Surv(RFS_months,RFS_status) ~ SNF,data = luminal)
+p = ggsurvplot(km,pval = T,pval.method = T,palette =  c(""#2378B3"",  ""#1CB038"",  ""#F8A900"", ""#D5271A""),
+               break.x.by = 12)
+ggsave(p$plot,filename = paste0(""surplot_SNF_RFS.pdf""),height = 6,width = 6)
+graph2ppt(p$plot,file =paste0(""surplot_SNF_RFS.pptx""),height = 6,width = 6)
+
+
+
+############################ COX ---------
+## COX forest --------
+colnames(luminal)[colnames(luminal) == ""辅助内分泌治疗""  ] =""EndoTherapy"" 
+luminal = luminal[luminal$EndoTherapy %in% c(""Yes"",""No""),]  
+
+colnames(luminal)[colnames(luminal) == ""淋巴结转移数"" ] =""LN_count"" 
+luminal$LN_count = as.numeric(luminal$LN_count)
+
+colnames(luminal)[colnames(luminal) == ""辅助化疗"" ] =""ChemoTherapy"" 
+luminal = luminal[luminal$ChemoTherapy %in% c(""Yes"",""No""),]
+
+colnames(luminal)[colnames(luminal) == ""肿瘤大小.cm."" ] =""TumorSize"" 
+luminal$TumorSize = as.numeric(luminal$TumorSize)
+
+colnames(luminal)[colnames(luminal) == ""组织学分级"" ] =""Grade"" 
+luminal$Grade[luminal$Grade  %in% c(""Unknown"",""Unknown（组织学类型不明）"")] = ""Unknown""
+luminal$Grade[luminal$Grade == ""1.5""] = ""2""
+luminal$Grade[luminal$Grade == ""2.5""] = ""3""
+luminal$Grade = as.numeric(luminal$Grade)
+
+colnames(luminal)[colnames(luminal) == ""PAM50_classifier"" ] =""PAM50"" 
+luminal$PAM50[! luminal$PAM50 %in% c(""LumA"",""LumB"")] = ""nonLum""
+#luminal$PAM50 = factor(luminal$PAM50,levels = c(""LumA"",""LumB"",""Her2"",""Basal"",""Normal""))
+luminal$PAM50 = factor(luminal$PAM50,levels = c(""LumA"",""LumB"",""nonLum""))
+
+luminal = luminal[luminal$EndoTherapy == ""Yes"",]
+
+res.cox = coxph(Surv(DMFS_months,DMFS_status)~ SNF + LN_count + ChemoTherapy+TumorSize+Grade+PAM50 , data = luminal)
+res.cox = coxph(Surv(DMFS_months,DMFS_status)~ SNF + LN_count + ChemoTherapy+TumorSize+Grade , data = luminal)
+res.cox = summary(res.cox)
+
+tmp1 = res.cox[[""coefficients""]]
+tmp2 = res.cox[[""conf.int""]]
+cox.res = as.data.frame(cbind(tmp1[,c(2,5)],tmp2[,3:4]))
+colnames(cox.res) = c(""HR"",""p.val"",""lower.95"",""upper.95"")
+
+write.csv(cox.res,file ='./SNF_TS_LN_CT_COX4plot_pickET.csv' )
+######### forest plot
+cox.res$p.val = round(cox.res$p.val, 2)
+cox.res$factor = rownames(cox.res)
+#pdf(""SNF_TS_LN_CT_multiCOX_forest_mergeNoLum_pickET.pdf"",width = 8,height = 4)
+pdf(""SNF_TS_LN_CT_multiCOX_forest_pickET.pdf"",width = 8,height = 4)
+forestplot(labeltext= as.matrix(cox.res[,c(5,2)]), graph.pos=2,
+           mean=c(cox.res$HR),
+           lower=c(cox.res$lower.95), upper=c(cox.res$upper.95),
+           txt_gp=fpTxtGp(label=gpar(cex=1.25),
+                          ticks=gpar(cex=1.1),
+                          xlab=gpar(cex = 1.2),
+                          title=gpar(cex = 1.2)),
+           col=fpColors(box=""#1c61b6"", lines=""#1c61b6"", zero = ""gray50""), 
+           zero=1, 
+           cex=0.5, lineheight = ""auto"",
+           colgap=unit(8,""mm""),
+           lwd.ci=2, boxsize=0.3, 
+           ci.vertices=TRUE,
+           ci.vertices.height = 0.1) 
+
+graph2ppt(file = ""SNF_TS_LN_CT_multiCOX_forest_pickET.ppt"",width = 8,height = 4)
+dev.off()
+","R"
+"Receptor","yifanzhou330/Luminal-SNF","Fig3_and_related_ED_Fig/Clinical_comparison.R",".R","13886","301","rm(list = ls()) ; graphics.off()
+library(ggplot2)
+library(export)
+library(plyr)
+
+load(""/CBCGA_HRposHER2neg351_WES_RNAseq_CNV_Metab_Protein_20220829.Rdata"")
+luminal$SNF = paste0(""SNF"",SNF_Cluster[rownames(luminal)])
+
+# age-------
+luminal$Age_2[luminal$Age <= 40] = ""Age:<=40""
+luminal$Age_2[luminal$Age > 40] = ""Age:>40""
+cluster.merge4plot = as.data.frame(table(luminal$SNF, luminal$Age_2))
+colnames(cluster.merge4plot)[1:2] = c(""SNF"",""Age"")
+
+cluster.merge4plot2 = ddply(cluster.merge4plot,""SNF"", transform,
+                            percent = Freq / sum(Freq) *100)
+cluster.merge4plot2$Age = factor(cluster.merge4plot2$Age,levels = c(""Age:<=40"",""Age:>40""))
+#color = c(""#DBBEBE"", ""#C88E8D"", ""#B75C5B"", ""#A52B29"")
+color = c(""#DBBEBE"",  ""#A52B29"")
+names(color) = c(""Age:<=40"",""Age:>40"")
+set.seed(123)
+tmp = fisher.test(table(luminal$SNF, luminal$Age_2),simulate.p.value=T)
+p = ggplot(cluster.merge4plot2, aes(x=SNF, y = percent, fill = Age )) + geom_bar(stat = ""identity"") +
+  scale_fill_manual(values = color)+
+  theme(panel.border = element_rect(fill=NA, colour = ""black"", size=0.75), 
+        axis.ticks = element_line(size = 0.75),
+        axis.text.x=element_text(size = 7),
+        axis.text.y=element_text(size = 7),
+        axis.title.x = element_text(size = 7),
+        axis.title.y = element_text(size = 7),
+        panel.background = element_blank(),
+        panel.grid.major = element_blank(), 
+        panel.grid.minor = element_blank() 
+  ) +
+  ggtitle(paste0(""p = "",round(tmp[[""p.value""]],5)) ) 
+ggsave(p,filename =""./SNF_Age_4plot_percent_barplot_2.pdf"")
+
+
+# Ki_67-----
+luminal$Ki_67_revise = luminal$Ki67阳性百分比 
+luminal_tmp = luminal[!is.na(luminal$Ki_67_revise),]
+luminal_tmp$Ki_67_revise[luminal_tmp$Ki67阳性百分比 < 15] = ""ki_67:low""
+luminal_tmp$Ki_67_revise[luminal_tmp$Ki67阳性百分比 < 30 & luminal_tmp$Ki67阳性百分比 >= 15 ] = ""ki_67:middle""
+luminal_tmp$Ki_67_revise[luminal_tmp$Ki67阳性百分比 >= 30] = ""ki_67:high""
+cluster.merge4plot = as.data.frame(table(luminal_tmp$SNF, luminal_tmp$Ki_67_revise))
+colnames(cluster.merge4plot)[1:2] = c(""SNF"",""Ki_67"")
+
+cluster.merge4plot2 = ddply(cluster.merge4plot,""SNF"", transform,
+                            percent = Freq / sum(Freq) *100)
+cluster.merge4plot2$Ki_67 = factor(cluster.merge4plot2$Ki_67,levels = c(""ki_67:low"",""ki_67:middle"",""ki_67:high""))
+color = c(""#EDD4BC"", ""#EFA463"",""#F27900"")
+names(color) = c(""ki_67:low"",""ki_67:middle"",""ki_67:high"")
+set.seed(123)
+tmp = fisher.test(table(luminal$SNF, luminal$Ki_67_revise),simulate.p.value=T)
+p = ggplot(cluster.merge4plot2, aes(x=SNF, y = percent, fill = Ki_67 )) + geom_bar(stat = ""identity"") +
+  scale_fill_manual(values = color)+
+  theme(panel.border = element_rect(fill=NA, colour = ""black"", size=0.75), 
+        axis.ticks = element_line(size = 0.75),
+        axis.text.x=element_text(size = 7),
+        axis.text.y=element_text(size = 7),
+        axis.title.x = element_text(size = 7),
+        axis.title.y = element_text(size = 7),
+        panel.background = element_blank(),
+        panel.grid.major = element_blank(),
+        panel.grid.minor = element_blank() 
+  ) +
+  ggtitle(paste0(""p = "",round(tmp[[""p.value""]],5)) ) 
+ggsave(p,filename =""./SNF_Ki_67_4plot_percent_barplot.pdf"")
+#graph2ppt(p, file = ""SNF_Ki_67_4plot_percent_barplot.ppt"")
+
+# Menopause_status-----
+luminal_tmp = luminal[luminal$绝经状态 %in% c(""Yes"",""No""),]
+luminal_tmp$绝经状态[luminal_tmp$绝经状态 %in% c(""Yes"")] = ""TRUE""
+luminal_tmp$绝经状态[luminal_tmp$绝经状态 %in% c(""No"")] = ""FALSE""
+
+cluster.merge4plot = as.data.frame(table(luminal_tmp$SNF, luminal_tmp$绝经状态))
+colnames(cluster.merge4plot)[1:2] = c(""SNF"",""Menopause_status"")
+
+cluster.merge4plot2 = ddply(cluster.merge4plot,""SNF"", transform,
+                            percent = Freq / sum(Freq) *100)
+cluster.merge4plot2$Menopause_status = factor(cluster.merge4plot2$Menopause_status,levels = c(""TRUE"",""FALSE""))
+color = c(""#7F7E7E"", ""#C4C4C4"")
+names(color) = c(""TRUE"",""FALSE"")
+set.seed(123)
+tmp = fisher.test(table(luminal_tmp$SNF, luminal_tmp$绝经状态),simulate.p.value=T)
+p = ggplot(cluster.merge4plot2, aes(x=SNF, y = percent, fill = Menopause_status )) + geom_bar(stat = ""identity"") +
+  scale_fill_manual(values = color)+
+  theme(panel.border = element_rect(fill=NA, colour = ""black"", size=0.75), 
+        axis.ticks = element_line(size = 0.75),
+        axis.text.x=element_text(size = 7),
+        axis.text.y=element_text(size = 7),
+        axis.title.x = element_text(size = 7),
+        axis.title.y = element_text(size = 7),
+        panel.background = element_blank(),
+        panel.grid.major = element_blank(), 
+        panel.grid.minor = element_blank() 
+  ) +
+  ggtitle(paste0(""p = "",round(tmp[[""p.value""]],5)) ) 
+ggsave(p,filename =""./SNF_Menopause_status_4plot_percent_barplot.pdf"")
+#graph2ppt(p, file = ""SNF_Menopause_status_4plot_percent_barplot.ppt"")
+
+# LVI-----
+luminal_tmp = luminal[luminal$脉管癌栓 %in% c(""Yes"",""No""),]
+luminal_tmp$脉管癌栓[luminal_tmp$脉管癌栓 %in% ""Yes""] = ""TRUE""
+luminal_tmp$脉管癌栓[luminal_tmp$脉管癌栓 %in% ""No""] = ""FALSE""
+cluster.merge4plot = as.data.frame(table(luminal_tmp$SNF, luminal_tmp$脉管癌栓))
+colnames(cluster.merge4plot)[1:2] = c(""SNF"",""LVI"")
+
+cluster.merge4plot2 = ddply(cluster.merge4plot,""SNF"", transform,
+                            percent = Freq / sum(Freq) *100)
+cluster.merge4plot2$LVI = factor(cluster.merge4plot2$LVI,levels = c(""TRUE"",""FALSE""))
+color = c(""#7F7E7E"", ""#C4C4C4"")
+names(color) = c(""TRUE"",""FALSE"")
+set.seed(123)
+tmp = fisher.test(table(luminal_tmp$SNF, luminal_tmp$脉管癌栓),simulate.p.value=T)
+p = ggplot(cluster.merge4plot2, aes(x=SNF, y = percent, fill = LVI )) + geom_bar(stat = ""identity"") +
+  scale_fill_manual(values = color)+
+  theme(panel.border = element_rect(fill=NA, colour = ""black"", size=0.75), 
+        axis.ticks = element_line(size = 0.75),
+        axis.text.x=element_text(size = 7),
+        axis.text.y=element_text(size = 7),
+        axis.title.x = element_text(size = 7),
+        axis.title.y = element_text(size = 7),
+        panel.background = element_blank(),
+        panel.grid.major = element_blank(), 
+        panel.grid.minor = element_blank() 
+  ) +
+  ggtitle(paste0(""p = "",round(tmp[[""p.value""]],5)) ) 
+ggsave(p,filename =""./SNF_LVI_4plot_percent_barplot.pdf"")
+#graph2ppt(p, file = ""SNF_LVI_4plot_percent_barplot.ppt"")
+
+
+# Adjuvant_chemotherapy-----
+luminal_tmp = luminal[luminal$辅助化疗 %in% c(""Yes"",""No""),]
+luminal_tmp$辅助化疗[luminal_tmp$辅助化疗 %in% ""Yes""] = ""TRUE""
+luminal_tmp$辅助化疗[luminal_tmp$辅助化疗 %in% ""No""] = ""FALSE""
+cluster.merge4plot = as.data.frame(table(luminal_tmp$SNF, luminal_tmp$辅助化疗))
+colnames(cluster.merge4plot)[1:2] = c(""SNF"",""Adjuvant_chemotherapy"")
+
+cluster.merge4plot2 = ddply(cluster.merge4plot,""SNF"", transform,
+                            percent = Freq / sum(Freq) *100)
+cluster.merge4plot2$Adjuvant_chemotherapy = factor(cluster.merge4plot2$Adjuvant_chemotherapy,levels = c(""TRUE"",""FALSE""))
+color = c(""#7F7E7E"", ""#C4C4C4"")
+names(color) = c(""TRUE"",""FALSE"")
+set.seed(123)
+tmp = fisher.test(table(luminal_tmp$SNF, luminal_tmp$辅助化疗),simulate.p.value=T)
+p = ggplot(cluster.merge4plot2, aes(x=SNF, y = percent, fill = Adjuvant_chemotherapy )) + geom_bar(stat = ""identity"") +
+  scale_fill_manual(values = color)+
+  theme(panel.border = element_rect(fill=NA, colour = ""black"", size=0.75),
+        axis.ticks = element_line(size = 0.75),
+        axis.text.x=element_text(size = 7),
+        axis.text.y=element_text(size = 7),
+        axis.title.x = element_text(size = 7),
+        axis.title.y = element_text(size = 7),
+        panel.background = element_blank(),
+        panel.grid.major = element_blank(), 
+        panel.grid.minor = element_blank() 
+  ) +
+  ggtitle(paste0(""p = "",round(tmp[[""p.value""]],5)) ) 
+ggsave(p,filename =""./SNF_Adjuvant_chemotherapy_4plot_percent_barplot.pdf"")
+#graph2ppt(p, file = ""SNF_Adjuvant_chemotherapy_4plot_percent_barplot.ppt"")
+
+
+# Radiotherapy-----
+luminal_tmp = luminal[luminal$辅助放疗 %in% c(""Yes"",""No""),]
+luminal_tmp$辅助放疗[luminal_tmp$辅助放疗 %in% ""Yes""] = ""TRUE""
+luminal_tmp$辅助放疗[luminal_tmp$辅助放疗 %in% ""No""] = ""FALSE""
+cluster.merge4plot = as.data.frame(table(luminal_tmp$SNF, luminal_tmp$辅助放疗))
+colnames(cluster.merge4plot)[1:2] = c(""SNF"",""Radiotherapy"")
+
+cluster.merge4plot2 = ddply(cluster.merge4plot,""SNF"", transform,
+                            percent = Freq / sum(Freq) *100)
+cluster.merge4plot2$Radiotherapy = factor(cluster.merge4plot2$Radiotherapy,levels = c(""TRUE"",""FALSE""))
+color = c(""#7F7E7E"", ""#C4C4C4"")
+names(color) = c(""TRUE"",""FALSE"")
+set.seed(123)
+tmp = fisher.test(table(luminal_tmp$SNF, luminal_tmp$辅助放疗),simulate.p.value=T)
+p = ggplot(cluster.merge4plot2, aes(x=SNF, y = percent, fill = Radiotherapy )) + geom_bar(stat = ""identity"") +
+  scale_fill_manual(values = color)+
+  theme(panel.border = element_rect(fill=NA, colour = ""black"", size=0.75), 
+        axis.ticks = element_line(size = 0.75),
+        axis.text.x=element_text(size = 7),
+        axis.text.y=element_text(size = 7),
+        axis.title.x = element_text(size = 7),
+        axis.title.y = element_text(size = 7),
+        panel.background = element_blank(),
+        panel.grid.major = element_blank(),
+        panel.grid.minor = element_blank() 
+  ) +
+  ggtitle(paste0(""p = "",round(tmp[[""p.value""]],5)) ) 
+  ggsave(p,filename =""./SNF_Radiotherapy_4plot_percent_barplot.pdf"")
+#graph2ppt(p, file = ""SNF_Radiotherapy_4plot_percent_barplot.ppt"")
+
+
+
+# Grade -----
+luminal_tmp = luminal[luminal$组织学分级 != ""Unknown"",]
+luminal_tmp$Grade = luminal_tmp$组织学分级
+luminal_tmp$Grade[luminal_tmp$Grade %in% ""2.5""] = 3
+luminal_tmp$Grade[luminal_tmp$Grade %in% ""1.5""] = 2
+luminal_tmp$Grade = as.numeric(luminal_tmp$Grade)
+luminal_tmp$Grade[luminal_tmp$Grade > 2] = "">G2""
+luminal_tmp$Grade[luminal_tmp$Grade != "">G2"" ] = ""<=G2""
+
+cluster.merge4plot = as.data.frame(table(luminal_tmp$SNF, luminal_tmp$Grade))
+colnames(cluster.merge4plot)[1:2] = c(""SNF"",""Grade"")
+
+cluster.merge4plot2 = ddply(cluster.merge4plot,""SNF"", transform,
+                            percent = Freq / sum(Freq) *100)
+cluster.merge4plot2$Grade = factor(cluster.merge4plot2$Grade,levels = c(""<=G2"","">G2""))
+color = c(""#BDD7EE"", ""#2E75B6"")
+names(color) = c(""<=G2"","">G2"")
+set.seed(123)
+tmp = fisher.test(table(luminal_tmp$SNF, luminal_tmp$Grade),simulate.p.value=T)
+p = ggplot(cluster.merge4plot2, aes(x=SNF, y = percent, fill = Grade )) + geom_bar(stat = ""identity"") +
+  scale_fill_manual(values = color)+
+  theme(panel.border = element_rect(fill=NA, colour = ""black"", size=0.75), 
+        axis.ticks = element_line(size = 0.75),
+        axis.text.x=element_text(size = 7),
+        axis.text.y=element_text(size = 7),
+        axis.title.x = element_text(size = 7),
+        axis.title.y = element_text(size = 7),
+        panel.background = element_blank(),
+        panel.grid.major = element_blank(), 
+        panel.grid.minor = element_blank() 
+  ) +
+  ggtitle(paste0(""p = "",round(tmp[[""p.value""]],5)) ) 
+  ggsave(p,filename =""./SNF_Grade_4plot_percent_barplot.pdf"")
+#graph2ppt(p, file = ""SNF_Grade_4plot_percent_barplot.ppt"")
+
+  
+# pT -----
+luminal$pT = luminal$pT.仅计算浸润成分.不计原位癌.
+luminal$pT[luminal$pT %in% c(""pT1a"", ""pT1b"", ""pT1c"", ""pT1mi"")] = ""pT1""
+cluster.merge4plot = as.data.frame(table(luminal$SNF, luminal$pT))
+colnames(cluster.merge4plot)[1:2] = c(""SNF"",""pT"")
+
+cluster.merge4plot2 = ddply(cluster.merge4plot,""SNF"", transform,
+                            percent = Freq / sum(Freq) *100)
+cluster.merge4plot2$pT = factor(cluster.merge4plot2$pT,levels = c(""pT1"" , ""pT2"" ,""pT3"" ))
+color = c(""#DBBEBE"", ""#C27978"", ""#A52B29"")
+names(color) = c(""pT1"" , ""pT2"" ,""pT3"" )
+set.seed(123)
+tmp = fisher.test(table(luminal$SNF, luminal$pT),simulate.p.value=T)
+p = ggplot(cluster.merge4plot2, aes(x=SNF, y = percent, fill = pT )) + geom_bar(stat = ""identity"") +
+  scale_fill_manual(values = color)+
+  theme(panel.border = element_rect(fill=NA, colour = ""black"", size=0.75), 
+        axis.ticks = element_line(size = 0.75),
+        axis.text.x=element_text(size = 7),
+        axis.text.y=element_text(size = 7),
+        axis.title.x = element_text(size = 7),
+        axis.title.y = element_text(size = 7),
+        panel.background = element_blank(),
+        panel.grid.major = element_blank(), 
+        panel.grid.minor = element_blank() 
+  ) +
+  ggtitle(paste0(""p = "",round(tmp[[""p.value""]],5)) ) 
+  ggsave(p,filename =""./SNF_pT_4plot_percent_barplot.pdf"")
+#graph2ppt(p, file = ""SNF_pT_4plot_percent_barplot.ppt"")
+
+
+# pN -----
+cluster.merge4plot = as.data.frame(table(luminal$SNF, luminal$pN))
+colnames(cluster.merge4plot)[1:2] = c(""SNF"",""pN"")
+
+cluster.merge4plot2 = ddply(cluster.merge4plot,""SNF"", transform,
+                            percent = Freq / sum(Freq) *100)
+cluster.merge4plot2$pN = factor(cluster.merge4plot2$pN,levels = c(""pN0"" , ""pN1"", ""pN2"", ""pN3"" ))
+color = c(""#E7DCE6"", ""#E0C7DC"", ""#D8B4D0"",  ""#D1A1C7"")
+names(color) = c(""pN0"" , ""pN1"", ""pN2"", ""pN3"" )
+set.seed(123)
+tmp = fisher.test(table(luminal$SNF, luminal$pN),simulate.p.value=T)
+p = ggplot(cluster.merge4plot2, aes(x=SNF, y = percent, fill = pN )) + geom_bar(stat = ""identity"") +
+  scale_fill_manual(values = color)+
+  theme(panel.border = element_rect(fill=NA, colour = ""black"", size=0.75), 
+        axis.ticks = element_line(size = 0.75),
+        axis.text.x=element_text(size = 7),
+        axis.text.y=element_text(size = 7),
+        axis.title.x = element_text(size = 7),
+        axis.title.y = element_text(size = 7),
+        panel.background = element_blank(),
+        panel.grid.major = element_blank(), 
+        panel.grid.minor = element_blank() 
+  ) +
+  ggtitle(paste0(""p = "",round(tmp[[""p.value""]],5)) ) 
+ggsave(p,filename = ""./SNF_pN_4plot_percent_barplot.pdf"")
+#graph2ppt(p, file = ""SNF_pN_4plot_percent_barplot.ppt"")
+
+
+
+
+
+
+
+
+
+
+
+
+
+","R"
+"Receptor","yifanzhou330/Luminal-SNF","Fig6_and_related_ED_Fig/Phos_analysis.R",".R","5258","61","rm(list = ls()) ; graphics.off()
+setwd(""/Users/ZYF/Desktop/研究生科研/研一/lumianl平台建立/生信分析/31.NG二修/分型外推/SVM_RF/result/inferSNF_res_EachSNFfixed_V2/CPTAC_2022/DEPhos"")
+library(clusterProfiler)
+test_summary = openxlsx:::read.xlsx(""/Users/ZYF/Desktop/研究生科研/研一/lumianl平台建立/生信分析/31.NG二修/分型外推/SVM_RF/rawdata/inferSNF_res_EachSNFfixed_V2/RawNum100_CorThresh0.75_RF_scale_fold0_CPTAC.xlsx"")
+rownames(test_summary) = test_summary$X1
+phos = data.table::fread(""/Users/ZYF/Desktop/bioinformatics/bioinformatics_resource/外部组学队列/CPTAC/CPTAC 2020/S060_Breast_Cancer_data_freeze_GCTfiles_v5.4-public/prosp-brca-v5.4-public-phosphoproteome-ratio-norm-NArm.gct.csv"",
+                         data.table = F)
+rownames(phos) = phos[,1]
+phos_annot = phos[,c(1:22)]
+phos = phos[,-c(1:22)]
+
+tmp = intersect(rownames(test_summary),colnames(phos))
+phos = phos[,tmp]
+test_summary = test_summary[tmp,]
+
+# tmp = apply(phos,1,function(x){ sum(is.na(x))/ncol(phos) < 0.8 })
+# phos = phos[tmp,]
+
+########
+phos_SNF4 = phos[,rownames(test_summary[test_summary$inferSNF == ""SNF4"",])]
+phos_Others = phos[,rownames(test_summary[test_summary$inferSNF != ""SNF4"",])]
+
+wil.res = as.data.frame(matrix(nrow = nrow(phos_SNF4), ncol = 5))
+rownames(wil.res) = rownames(phos_SNF4)
+colnames(wil.res) = c(""SNF4"",""Others"",""Diff"",""pvalue"",""padj"")
+
+for ( k in rownames(wil.res)){
+  mat = data.frame(gene = c(as.numeric(phos_SNF4[k,]), as.numeric(phos_Others[k,]) ) ,
+                   group = c(rep(""SNF4"", ncol(phos_SNF4)), rep(""Others"",ncol(phos_Others)))
+  )
+  mat = mat[!is.na(mat$gene),]
+  if(length(unique(mat$group)) != 2){ next }
+  
+  res = wilcox.test(gene~group, data = mat,exact = FALSE )
+  wil.res[k,""pvalue""] = res$p.value
+  wil.res[k,""SNF4""] = mean(mat$gene[mat$group == ""SNF4""])
+  wil.res[k,""Others""] = mean(mat$gene[mat$group == ""Others""])
+}
+wil.res$Diff = wil.res[,""SNF4""] - wil.res[,""Others""]
+wil.res$padj = p.adjust(wil.res$pvalue,method = ""fdr"")
+
+wil.res = cbind(wil.res, phos_annot[rownames(wil.res),])
+write.csv(wil.res, file = paste0(""./DEPhos.Wilcox.CPTAC.SNF4VsOthers.csv""))
+
+# tmp1 = c(""TGFBR3L"", ""MERTK"", ""ACVR1C"", ""CSF1R"", ""HJV"", ""EFNA3"", ""EFNA4"", ""EFNB3"", ""EGFR"", ""EPHA2"", ""ENG"", ""EPHA1"", ""EPHA3"", ""EPHA4"", ""EPHA5"", ""EPHA7"", ""EPHA8"",
+#         ""EPHB1"", ""EPHB2"", ""EPHB3"", ""EPHB4"", ""EPHB6"", ""ERBB2"", ""ERBB3"", ""ERBB4"", ""EFEMP1"", ""FGFR1"", ""FGFR3"", ""FGFR2"", ""FGFR4"", ""FLT1"", ""FLT3"", ""FLT4"", ""ALK"", ""SOSTDC1"", ""AMHR2"", ""EPHA10"", ""EPHA6"", ""IGF1R"", ""IGF2R"", ""INSR"", ""INSRR"", ""KDR"", ""KIT"", ""LTBP1"", ""LTK"", ""MET"", ""MST1R"", ""MUSK"", ""NTRK1"", ""NTRK2"", ""NTRK3"", ""ROR2"", ""DDR2"", ""CRIM1"", ""PDGFRA"", ""PDGFRL"", ""PDGFRB"", ""FGFRL1"", ""AXL"", ""RET"", ""ROS1"", ""BMPR1A"", ""BMPR1B"", ""BMPR2"", ""TEK"", ""TGFBR1"", ""TGFBR2"", ""TGFBR3"", ""TIE1"", ""TYRO3"", ""DDR1"", ""LTBP4"", ""NRP2"", ""NRP1"", ""ACVR1"", ""ACVR1B"", ""ACVR2A"", ""ACVR2B"", ""ACVRL1"")
+# tmp2 = c(""MERTK"", ""CSF1R"", ""EFNA3"", ""EFNA4"", ""EFNB3"", ""EGFR"", ""EPHA2"", ""EPHA1"", ""EPHA3"", ""EPHA4"", ""EPHA5"", ""EPHA7"", ""EPHA8"", ""EPHB1"", ""EPHB2"", ""EPHB3"", ""EPHB4"", ""EPHB6"", ""ERBB2"", ""ERBB3"", ""ERBB4"", ""EFEMP1"", ""FGFR1"", ""FGFR3"", ""FGFR2"", ""FGFR4"", ""FLT1"", ""FLT3"", ""FLT4"", ""ALK"", ""EPHA10"", ""EPHA6"", ""IGF1R"", ""IGF2R"", ""INSR"", ""INSRR"", ""KDR"", ""KIT"", ""LTK"", ""MET"", ""MST1R"", ""MUSK"", ""NTRK1"", ""NTRK2"", ""NTRK3"", ""ROR2"", ""DDR2"", ""CRIM1"", ""PDGFRA"", ""PDGFRL"", ""PDGFRB"", ""FGFRL1"", ""AXL"", ""RET"", ""ROS1"", ""TEK"", ""TIE1"", ""TYRO3"", ""DDR1"", ""NRP2"", ""NRP1"")
+# tmp3 = c(""HIPK3"", ""TNK2"", ""TESK2"", ""MERTK"", ""CAMKK2"", ""CLK1"", ""CLK2"", ""CLK3"", ""CSF1R"", ""CSK"", ""HIPK4"", ""DYRK1A"", ""EFNA3"", ""EFNA4"", ""EFNB3"", ""EGFR"", ""EPHA2"", ""EPHA1"", ""EPHA3"", ""EPHA4"", ""EPHA5"", ""EPHA7"", ""EPHA8"", ""EPHB1"", ""EPHB2"", ""HIPK1"", ""EPHB3"", ""EPHB4"", ""EPHB6"", ""ERBB2"", ""ERBB3"", ""ERBB4"", ""PTK2B"", ""EFEMP1"", ""FER"", ""FES"", ""FGFR1"", ""FGFR3"", ""FGFR2"", ""FGFR4"", ""FGR"", ""FLT1"", ""FLT3"", ""FLT4"", ""ALK"", ""FRK"", ""ABL1"", ""FYN"", ""DSTYK"", ""ABL2"", ""EPHA10"", ""EPHA6"", ""HIPK2"", ""HCK"", ""IGF1R"", ""IGF2R"", ""INSR"", ""INSRR"", ""ITK"", ""JAK1"", ""JAK2"", ""JAK3"", ""KDR"", ""KIT"", ""LCK"", ""LTK"",
+#          ""LYN"", ""MATK"", ""MET"", ""MST1R"", ""MUSK"", ""NEK1"", ""NTRK1"", ""NTRK2"", ""NTRK3"", ""ROR2"", ""DDR2"", ""WEE2"", ""CRIM1"", ""PDGFRA"", ""PDGFRL"", ""PDGFRB"", ""FGFRL1"", ""STYK1"", ""AXL"", ""PRKCD"", ""MAP2K1"", ""MAP2K2"", ""MAP2K3"", ""MAP2K5"", ""MAP2K6"", ""MAP2K7"", ""EIF2AK2"", ""CLK4"", ""SCYL1"", ""PTK2"", ""PTK6"", ""RET"", ""ROS1"", ""BLK"", ""MAP2K4"", ""SLA"", ""BMX"", ""SRC"", ""SRMS"", ""SYK"", ""BTK"", ""TEC"", ""TEK"", ""TESK1"", ""TIE1"", ""TTK"", ""TTN"", ""TXK"", ""TYK2"", ""TYRO3"", ""WEE1"", ""YES1"", ""ZAP70"", ""DDR1"", ""PEAK1"", ""DYRK3"", ""DYRK2"", ""TTBK1"", ""STK16"", ""TNK1"", ""RIPK2"", ""DYRK4"", ""NRP2"", ""NRP1"", ""BAZ1B"", ""PKDCC"", ""DYRK1B"", ""AATK"", ""MELK"")
+# tmp = union(union(tmp1,tmp2),tmp3)
+gmt = read.gmt(""/Users/ZYF/c5.go.mf.v7.4.symbols.gmt"")
+RTK = gmt[gmt$term == ""GOMF_TRANSMEMBRANE_RECEPTOR_PROTEIN_KINASE_ACTIVITY"",]
+rownames(RTK) = RTK$gene
+kinase = openxlsx::read.xlsx(""/Users/ZYF/Desktop/研究生科研/研一/lumianl平台建立/生信分析/30.NG修改/分型外推/CPTAC_2020/激酶kinase 参考文献：Driver Fusions and Their Implications in the Development and Treatment of Human Cancers.xlsx"",1)
+colnames(kinase) = kinase[1,] ; kinase = kinase[-1,] 
+RTK = intersect(rownames(RTK),kinase$Kinase)
+
+wil.res2 = wil.res[wil.res$GeneSymbol %in% RTK,]
+write.csv(wil.res2,file = ""DEPhos.RTK.kinase.Wilcox.CPTAC.SNF4VsOthers.csv"")
+
+","R"
+"Receptor","yifanzhou330/Luminal-SNF","Fig6_and_related_ED_Fig/RTK_specificGene_SNF1_4.R",".R","2953","68","rm(list = ls()) ; graphics.off()
+setwd(""/Users/ZYF/Desktop/研究生科研/研一/lumianl平台建立/生信分析/31.NG二修/分型外推/SVM_RF/result/inferSNF_res_EachSNFfixed_V2/TCGA/RawNum100_CorThresh0.75_RF_scale_fold0/RTK_specificGene_SNF1_4"")
+library(stringi)
+library(stringr)
+library(ggpubr)
+library(export)
+load(""/Users/ZYF/Desktop/研究生科研/研一/lumianl平台建立/生信分析/27.Fig重绘/rawdata/TCGA/TCGA_hg38_FPKM_Mut_hg19ascatCNV.Rdata"")
+SNF.file = ""RawNum100_CorThresh0.75_RF_scale_fold0_TCGA""
+SNFnew = openxlsx::read.xlsx(paste0(""/Users/ZYF/Desktop/研究生科研/研一/lumianl平台建立/生信分析/31.NG二修/分型外推/SVM_RF/rawdata/inferSNF_res_EachSNFfixed_V2/"",SNF.file,"".xlsx""),1,
+                             rowNames = T)
+SNF_Cluster = SNFnew$inferSNF ; names(SNF_Cluster) = rownames(SNFnew)
+
+
+gene = c(""EGFR"",""ALK"", ""KIT"", ""PDGFRA"",""MET"")
+for (i in gene){
+  #  RNA 
+  exp.fpkm.TT.log = log2(exp_fpkm_TCGA+1)
+  EGFR = as.data.frame(t(exp.fpkm.TT.log[i,]))
+  colnames(EGFR) = ""gene""
+  EGFR$SNF = SNF_Cluster[colnames(exp.fpkm.TT.log)]
+  EGFR$SNF[EGFR$SNF != ""SNF4""] = ""Others""
+  EGFR$SNF = factor(EGFR$SNF,levels = c(""SNF4"",""Others""))
+  #EGFR$SNF = factor(paste0(""SNF"",EGFR$SNF),levels = paste0(""SNF"",c(1:4)))
+  
+  shapiro.test(EGFR$gene)
+  bartlett.test(gene~SNF,data=EGFR)
+  
+  # p = ggviolin(EGFR,x = ""SNF"",y= ""gene"",fill = ""SNF"",
+  #              add = ""boxplot"", add.params = list(fill=""white""),
+  #              palette = c(color[4],""grey"") )+
+  #   #ylim(0,) +
+  #   stat_compare_means(label.y = max(EGFR$gene)+1) +
+  #   xlab(i) + 
+  p = ggboxplot(EGFR,x = ""SNF"",y= ""gene"",color = ""SNF"",
+               add = ""jitter"", 
+               palette = c(color[4],""grey"") )+
+    #ylim(0,) +
+    stat_compare_means(label.y = max(EGFR$gene)+1) +
+    xlab(i) 
+  ggsave(p,filename = paste0(""./TCGA_SNF4_Others_"",i,""_FPKMlog.pdf""), height = 4, width = 4)
+  graph2ppt(p,file = paste0(""./TCGA_SNF4_Others_"",i,""_FPKMlog.ppt""), height = 4, width = 4)
+}
+
+i = ""ALK""
+EGFR = as.data.frame(t(exp.fpkm.TT.log[i,]))
+colnames(EGFR) = ""gene""
+EGFR$SNF = SNF_Cluster[colnames(exp.fpkm.TT.log)]
+EGFR$SNF[EGFR$SNF != ""SNF4""] = ""Others""
+EGFR$SNF = factor(EGFR$SNF,levels = c(""SNF4"",""Others""))
+#EGFR$SNF = factor(paste0(""SNF"",EGFR$SNF),levels = paste0(""SNF"",c(1:4)))
+
+shapiro.test(EGFR$gene)
+bartlett.test(gene~SNF,data=EGFR)
+
+# p = ggviolin(EGFR,x = ""SNF"",y= ""gene"",fill = ""SNF"",
+#              add = ""boxplot"", add.params = list(fill=""white""),
+#              palette = c(color[4],""grey"") )+
+#   ylim(0,1) +
+#   stat_compare_means(label.y = 1) +
+#   xlab(i) + 
+p = ggboxplot(EGFR,x = ""SNF"",y= ""gene"",color = ""SNF"",
+             add = ""jitter"",
+             palette = c(color[4],""grey"") )+ ylim(0,1) +
+  stat_compare_means(label.y = 1) +
+  xlab(i) 
+ggsave(p,filename = paste0(""./TCGA_SNF4_Others_"",i,""_FPKMlog.pdf""), height = 4, width = 4)
+graph2ppt(p,file = paste0(""./TCGA_SNF4_Others_"",i,""_FPKMlog.ppt""), height = 4, width = 4)
+","R"
+"Receptor","yifanzhou330/Luminal-SNF","Fig6_and_related_ED_Fig/RTK.hp.mean.SNF4phos.R",".R","9398","185","rm(list = ls()) ; graphics.off()
+setwd(""/Users/ZYF/Desktop/研究生科研/研一/lumianl平台建立/生信分析/31.NG二修/分型外推/SVM_RF/result/inferSNF_res_EachSNFfixed_V2/CPTAC_2022/RTK.hp.mean.SNF4phos"")
+library(ggplot2)
+library(ggrepel)
+library(dplyr)
+library(ComplexHeatmap)
+library(export)
+library(clusterProfiler)
+library(Seurat)
+library(tidyverse)
+library(paletteer)
+library(circlize)
+library(RColorBrewer)
+########################### PART 1. FUSCC SNF RTK RNA mean hp ---------
+load(""/Users/ZYF/Desktop/研究生科研/研一/lumianl平台建立/生信分析/30.NG修改/rawdata/CBCGA_HRposHER2neg351_WES_RNAseq_CNV_Metab_Protein_20220829.Rdata"")
+gmt = read.gmt(""/Users/ZYF/c5.go.mf.v7.4.symbols.gmt"")
+RTK = gmt[gmt$term == ""GOMF_TRANSMEMBRANE_RECEPTOR_PROTEIN_KINASE_ACTIVITY"",]
+rownames(RTK) = RTK$gene
+kinase = openxlsx::read.xlsx(""/Users/ZYF/Desktop/研究生科研/研一/lumianl平台建立/生信分析/30.NG修改/分型外推/CPTAC_2020/激酶kinase 参考文献：Driver Fusions and Their Implications in the Development and Treatment of Human Cancers.xlsx"",1)
+colnames(kinase) = kinase[1,] ; kinase = kinase[-1,] 
+RTK = intersect(rownames(RTK),kinase$Kinase)
+
+exp.fpkm.TT.log = log2(exp.fpkm.TT +1 )
+RTK.hp = exp.fpkm.TT.log[RTK,]
+RTK.hp = as.data.frame(t(RTK.hp))
+RTK.hp$SNF = paste0(""SNF"",SNF_Cluster[rownames(RTK.hp)])
+RTK.hp.mean = aggregate(RTK.hp[,1:(ncol(RTK.hp)-1)],list(RTK.hp$SNF),mean)
+
+rownames(RTK.hp.mean) = RTK.hp.mean$Group.1 ; RTK.hp.mean = RTK.hp.mean[,-1]
+RTK.hp.mean  = as.data.frame(t(RTK.hp.mean))
+
+SNF4 = read.csv(""/Users/ZYF/Desktop/研究生科研/研一/lumianl平台建立/生信分析/30.NG修改/DEG_SNF/DEG.DESeq2.SNF4VsOthers.csv"",
+                row.names = 1)
+SNF4 = na.omit(SNF4)
+SNF4 = SNF4[intersect(rownames(RTK.hp.mean),rownames(SNF4)) , ]
+SNF4 = SNF4[order(SNF4$log2FoldChange,decreasing = T),]
+
+RTK.hp.mean = RTK.hp.mean[rownames(SNF4),]
+ord = rownames(RTK.hp.mean)
+RTK.hp.mean.FUSCCRNA.scale = t(scale(t(RTK.hp.mean)))
+
+########################### PART 2. Lumianl neo scRNA ---------
+rm(list = ls()[! ls() %in% c(""RTK.hp.mean.FUSCCRNA.scale"",""ord"")])
+load(""/Users/ZYF/Desktop/研究生科研/研一/lumianl平台建立/生信分析/30.NG修改/rawdata/Luminal_9sample_Cancer鉴定后.Rdata"")
+raw = raw[ , ! raw@meta.data$major_V3 %in% c(""Epithelial"",""Prolifer"",""Normal_epithelial"")] ; gc()
+raw@meta.data$major_V3[raw@meta.data$major_V3 %in% c(""CD4T"",""CD8T"",""ISG_T"",""NK"")] = ""TNK""
+# dplot <- DotPlot(object = raw, features = ord, split.by = ""SNF_subtype"",group.by = ""merge_Cell_marjor_V2"",
+#                  cols = rainbow(100))
+dplot <- DotPlot(object = raw, features = ord, group.by = ""major_V3"",
+                 cols = rainbow(100))
+
+
+ddata <- as.data.frame(dplot[[""data""]])
+df_spread <- tidyr::spread(ddata[, c(3,4,5)], id, avg.exp.scaled)
+rownames(df_spread) <- df_spread[, 1]
+df_spread <- df_spread[, -1]
+mat <- as.matrix(df_spread)
+mat <- na.omit(mat)
+mat = mat[,order(colnames(mat))]
+scRNA.mean.hp.scale = t(scale(t(mat)))
+scRNA.mean.hp.scale = scRNA.mean.hp.scale[ord,]
+# phet<- pheatmap(scRNA_mean, fontsize = 6, cellheight = 6, border_color = T,#annotation_row = anno_row,
+#                 fontsize_row = 6,cluster_cols = F,cluster_rows = F)
+# phet
+
+scRNA.mean.hp.scale.annot = data.frame(row.names = colnames(scRNA.mean.hp.scale),
+                                       CellType = str_split_fixed( colnames(scRNA.mean.hp.scale),""_"",2)[,1]#,
+                                       #SNF = str_split_fixed( colnames(scRNA.mean.hp.scale),""_"",2)[,2]
+)
+
+########################### PART 3. CPTAC SNF matched RTK phos SNF diff annot ---------
+rm(list = ls()[! ls() %in% c(""RTK.hp.mean.FUSCCRNA.scale"",""ord"",""scRNA.mean.hp.scale"",""scRNA.mean.hp.scale.annot"")])
+
+## input DEphos
+DEphos = read.csv(""/Users/ZYF/Desktop/研究生科研/研一/lumianl平台建立/生信分析/31.NG二修/分型外推/SVM_RF/result/inferSNF_res_EachSNFfixed_V2/CPTAC_2022/DEPhos/DEPhos.Wilcox.CPTAC.SNF4VsOthers.csv"",
+                  row.names = 1)
+DEphos$index = paste0(DEphos$geneSymbol,""_"",DEphos$variableSites)
+DEphos = na.omit(DEphos)
+## pick best phos site for SNF4 in given gene list 
+DEphos_f = DEphos[rownames(DEphos)[DEphos$geneSymbol %in% ord],]
+DEphos_f$index2 = paste0(DEphos_f$geneSymbol,DEphos_f$Diff)
+
+tmp <- DEphos_f %>% 
+  group_by(geneSymbol) %>%    
+  summarise(max = max(Diff,na.rm = T))
+tmp$index = paste0(tmp$geneSymbol,tmp$max)
+
+DEphos_f = DEphos_f[match(tmp$index,DEphos_f$index2),]
+rownames(DEphos_f) = DEphos_f$geneSymbol
+
+DEphos_f = DEphos_f[match(ord,rownames(DEphos_f)),]
+rownames(DEphos_f) = ord 
+rownames(DEphos_f)[na.omit(match(DEphos_f$geneSymbol,rownames(DEphos_f)))] = na.omit(DEphos_f[,""index""])
+DEphos_f$geneSymbol = ord
+
+DEphos_f = DEphos_f[match(rownames(RTK.hp.mean.FUSCCRNA.scale),DEphos_f$geneSymbol),]
+#RTK.hp.mean.FUSCCRNA.scale.annot = DEphos_f[rownames(RTK.hp.mean.FUSCCRNA.scale),""Diff""]
+RTK.hp.mean.FUSCCRNA.scale.annot = data.frame(row.names = rownames(DEphos_f),
+                                              Diff = DEphos_f[,""Diff""])
+########################### PART 4. merge & plot  ---------
+rm(list = ls()[!ls() %in% c(""RTK.hp.mean.FUSCCRNA.scale"",""RTK.hp.mean.FUSCCRNA.scale.annot"",
+                            ""scRNA.mean.hp.scale"",""scRNA.mean.hp.scale.annot"")])
+color_fun_SNF= c(""#2378B3"",  ""#1CB038"",  ""#F8A900"", ""#D5271A"")
+names(color_fun_SNF) = paste0(""SNF"",1:4)
+
+color_fun_CellType = as.character(paletteer_d(""ggsci::default_nejm"")[-4])
+names(color_fun_CellType) = c(""B"", ""CAF"", ""Endothelial"", ""Myeloid"", ""NormalEpi"", ""TNK"", ""Tumor"")
+
+col_fun_FUSCCRNA.scale = colorRamp2(c(-1.2,-0.6, 0 , 0.6,1.2), rev(brewer.pal(n = 5, name =""RdBu"")))
+col_fun_scRNA.scale = colorRamp2(c(-2,-1, 0 , 1,2), rev(brewer.pal(n = 5, name =""PiYG"")))
+color_fun_phos_diff = colorRamp2(c(-1.75, -0.875, 0 , 0.875,1.75), rev(brewer.pal(n = 5, name =""PRGn"")))
+
+annot = data.frame(row.names = paste0(""SNF"",1:4),SNF = paste0(""SNF"",1:4))
+
+
+p1 = Heatmap(RTK.hp.mean.FUSCCRNA.scale,
+             na_col = ""#eaeaea"",
+             #border = T, 
+             #show_row_names = F, 
+             show_column_names = F,
+             show_heatmap_legend = T, 
+             row_names_side = ""left"",
+             row_names_gp = gpar(fontsize = 7),
+             col = col_fun_FUSCCRNA.scale,
+             heatmap_legend_param = list(title = ""bulk scaled \n averaged expression""),
+             cluster_rows = FALSE, cluster_columns = FALSE,
+             top_annotation = HeatmapAnnotation(SNF = annot$SNF,
+                                                show_annotation_name = F,
+                                                annotation_height = 7,
+                                                simple_anno_size = unit(3, ""mm""),
+                                                #annotation_name_side = ""right"", annotation_name_gp = gpar(fontsize = 7),
+                                                col = list(SNF = color_fun_SNF)) )
+
+{
+  CelltypeOrd = c(""Tumor"",""Myeloid"",""TNK"",""B"",""Endothelial"",""CAF"")
+  scRNA.mean.hp.scale = scRNA.mean.hp.scale[,CelltypeOrd]
+  scRNA.mean.hp.scale.annot = data.frame( row.names = CelltypeOrd,CellType =  CelltypeOrd)
+}
+p2 = Heatmap(scRNA.mean.hp.scale,
+             na_col = ""#eaeaea"",
+             #border = T, 
+             show_row_names = F,  show_column_names = F,
+             #row_names_side = ""left"",
+             row_names_gp = gpar(fontsize = 7),
+             heatmap_legend_param = list(title = ""scRNA-seq scaled \n averaged expression""),
+             col = col_fun_scRNA.scale,
+             cluster_rows = FALSE, cluster_columns = FALSE,
+             column_split = factor(scRNA.mean.hp.scale.annot[,""CellType""],levels = c(""Tumor"",""NormalEpi"",""Myeloid"",""TNK"",""B"",""Endothelial"",""CAF"")),
+             column_gap = unit(2, ""mm""),
+             column_title_gp = gpar(fontsize = 7),
+             column_title_rot = 0,
+             width = 5,
+             top_annotation = HeatmapAnnotation(CellType = scRNA.mean.hp.scale.annot$CellType,
+                                                #SNF = scRNA.mean.hp.scale.annot$SNF,
+                                                show_annotation_name = F,
+                                                #annotation_name_side = ""right"", 
+                                                annotation_name_gp = gpar(fontsize = 7),
+                                                simple_anno_size = unit(3, ""mm""),
+                                                col = list(#SNF = color_fun_SNF,
+                                                  CellType = color_fun_CellType)) 
+) 
+p2
+p3 = pheatmap(RTK.hp.mean.FUSCCRNA.scale.annot, 
+              na_col = ""#eaeaea"",
+              show_colnames = F,
+              border_color = ""white"", 
+              display_numbers = T,
+              fontsize = 7,
+              cellwidth = 20,
+              col = color_fun_phos_diff,
+              heatmap_legend_param = list(title = ""Diff for SNF4""),
+              cluster_rows = FALSE, cluster_cols = FALSE)
+
+p = draw(p1+p2+p3, auto_adjust = F,
+         ht_gap = unit(c(5, 2), ""mm""))
+
+export::graph2pdf(p,file = ""RTK.hp.mean.SNF4phos.pdf"",width = 8, height =8)
+
+# ha1 = rowAnnotation(phos_diff = RTK.hp.mean.FUSCCRNA.scale.annot,
+#                     annotation_name_gp = gpar(fontsize = 7),
+#                     na_col = ""#eaeaea"",
+#                     col = list(phos_diff = color_fun_phos_diff) )
+# p = p1+ha1
+# draw(p1+ha1, auto_adjust = F)
+","R"
+"Receptor","yifanzhou330/Luminal-SNF","Fig5_and_related_ED_Fig/Integrated_immune_landscape.R",".R","8939","234","rm(list = ls())
+graphics.off()
+library(maftools)
+library(ComplexHeatmap)
+library(circlize)
+library(stringr)
+library(RColorBrewer)
+library(GSVA) ; library(GSEABase)
+
+load(""./CBCGA_HRposHER2neg351_WES_RNAseq_CNV_Metab_Protein_20220829.Rdata"")
+exp.fpkm.TT.log = log2(exp.fpkm.TT+1)
+
+luminal$SNF = paste0(""SNF"",SNF_Cluster[rownames(luminal)])
+luminal$SNF2orNot = luminal$SNF
+luminal$SNF2orNot[luminal$SNF2orNot != ""SNF2"" ]  = ""nonSNF2""
+
+#### annot bar -------
+{
+  res_estimate = data.table::fread(""./SNF_estimate_estimate_score.gct"")
+  res_estimate = as.data.frame(t(res_estimate[3:6,-1]))
+  colnames(res_estimate) = res_estimate[1,]
+  rownames(res_estimate) = res_estimate[,1]
+  res_estimate = res_estimate[-1,-1]
+  #res_estimate$StromalScore = as.numeric(scale(as.numeric(res_estimate$StromalScore)))
+  res_estimate$ImmuneScore = as.numeric(res_estimate$ImmuneScore)
+}
+luminal$ImmuneSignature = res_estimate[rownames(luminal),""ImmuneScore""]
+
+#
+{
+  CINscore = as.data.frame(matrix(nrow = length(SNF_Cluster), ncol = 2))
+  rownames(CINscore) = names(sort(SNF_Cluster))  ;colnames(CINscore) = c(""CINscore"",""SNF_All"")
+  CINscore$SNF_All = paste0(""SNF"",sort(SNF_Cluster))
+  for (i in names(sort(SNF_Cluster))){
+    tmp = cnv.alldata[,i]
+    tmp = tmp^2
+    CINscore[i,""CINscore""] = sum(tmp)
+  }
+  CINscore$CINscore = log2(CINscore$CINscore)
+}
+luminal$CINscore = CINscore[rownames(luminal),""CINscore""]
+
+#
+luminal$PAM50_PC_TCGA = factor(luminal$PAM50_classifier, levels =c( ""LumA"" ,""LumB"" ,""Her2"",""Basal"" ,""Normal""))
+
+#
+sTILS = openxlsx::read.xlsx(""./CBCGA_TILs_F.xlsx"",1)
+rownames(sTILS) = sTILS$PatientCode
+sTILS = sTILS[rownames(luminal),]
+luminal = cbind(luminal,sTILS)
+
+#
+Tls.sig = read.csv(""./ICB_signature.csv"",row.names = 1)
+Tls.sig = Tls.sig[rownames(luminal),]
+luminal = cbind(luminal,Tls.sig)
+
+#
+TCR_BCR = read.csv(""./cbcga_927sample_TCRBCR_220719.csv"",row.names = 1)
+  TCR_BCR = TCR_BCR[str_detect(TCR_BCR$ID , ""_T""),] 
+  TCR_BCR = TCR_BCR[substring(TCR_BCR$sample_id_detail,1,1) == ""B"",] #'KTWZ', 'RVAQ' 重复测了RNA
+  TCR_BCR = TCR_BCR[!is.na(TCR_BCR$ID),]
+  rownames(TCR_BCR) = TCR_BCR$Patient_ID
+TCR_BCR = TCR_BCR[rownames(luminal),]
+luminal = cbind(luminal,TCR_BCR)
+
+#
+protein_log2_NA = as.data.frame(matrix(NA,nrow = nrow(protein_log2), 
+                           ncol = length(setdiff(rownames(luminal),colnames(protein_log2))) ))
+rownames(protein_log2_NA) = rownames(protein_log2)
+colnames(protein_log2_NA) = setdiff(rownames(luminal),colnames(protein_log2))
+
+
+#### mat1 ---------
+tmp = c(""ImmuneSignature"",""TIS.signature"",""STAT1.signature"")
+mat1 = t(luminal[,tmp])
+mat1 = t(scale(t(mat1)))
+
+#### mat2 ---------
+tmp = c(""PDCD1"",""CD274"",""CTLA4"")
+mat2_RNA = exp.fpkm.TT.log[tmp, rownames(luminal)]
+mat2_RNA = t(scale(t(mat2_RNA)))
+rownames(mat2_RNA) = paste0(rownames(mat2_RNA),""_RNA"")  
+
+mat2 = mat2_RNA
+
+#### mat3 ---------
+tmp = c(""TRA_diversity"",""TRB_diversity"")
+mat3 =  t(luminal[,tmp])
+mat3 = t(scale(t(mat3)))
+
+#### mat4 ---------
+tmp = c(""CD8A"",""GZMA"",""PRF1"",""IDO1"")
+mat4_RNA = exp.fpkm.TT.log[tmp,rownames(luminal) ]
+mat4_RNA = t(scale(t(mat4_RNA)))
+rownames(mat4_RNA) = paste0(rownames(mat4_RNA),""_RNA"")
+
+mat4_Pro = protein_log2[intersect(rownames(protein_log2),tmp), ]
+mat4_Pro = t(scale(t(mat4_Pro)))
+mat4_Pro = cbind(mat4_Pro,protein_log2_NA[rownames(mat4_Pro),])
+mat4_Pro = mat4_Pro[,rownames(luminal) ]
+rownames(mat4_Pro) = paste0(rownames(mat4_Pro),""_Protein"")
+
+mat4_Polar = polar_metabolite_TT_MS2_log2[rownames(polar_metabolite_MS2_mapping)[polar_metabolite_MS2_mapping$metabolite_mapping_name == ""L-Kynurenine""],
+                                          rownames(luminal)]
+mat4_Polar = t(scale(t(mat4_Polar)))
+rownames(mat4_Polar) = ""L-Kynurenine""
+mat4 = rbind(mat4_RNA,mat4_Pro,mat4_Polar)
+
+mat4 = mat4[c(""CD8A_RNA"",""CD8A_Protein"",
+              ""GZMA_RNA"",""GZMA_Protein"",
+              ""PRF1_RNA"",""PRF1_Protein"",
+              ""IDO1_RNA"",""IDO1_Protein"",""L-Kynurenine""),]
+
+#### mat5 ---------
+geneset4imm<-read.csv(""./2015-CIBERSORT.csv"")
+list<- split(as.matrix(geneset4imm)[,1], geneset4imm[,2])
+
+gsva_matrix<- gsva(as.matrix(exp.fpkm.TT), list, method='ssgsea',kcdf='Gaussian',abs.ranking=TRUE)
+
+tmp = pheatmap::pheatmap(gsva_matrix,cluster_rows = T,cluster_cols = F,scale = ""row"")
+
+mat5 = gsva_matrix[tmp[[""tree_row""]][[""order""]], rownames(luminal)]
+mat5 = t(scale(t(mat5)))
+
+#### draw  ------
+## ord
+#ord = rownames(luminal[order(luminal$SNF2orNot,luminal$ImmuneSignature,decreasing = T),])
+ord = rownames(luminal[order(luminal$SNF,decreasing = F),])
+luminal = luminal[ord,]
+
+## annot 
+color_PAM50 = c(""#1D76BC"",""#76CFE6"",""#6E59A6"",""#E11D2E"",""#CDCFD0"")
+  names(color_PAM50) = c(""LumA"",""LumB"",""Her2"",""Basal"",""Normal"")
+names(color) = paste0(""SNF"",1:4)
+col_fun = colorRamp2(c(-2,-1,0 , 1,2), rev(brewer.pal(n = 5, name =""RdBu"")))
+
+col_fun_Isig <- colorRamp2(c(-0.5, 0.3 ,2), colorRampPalette(c(""white"", ""#08506F""))(20)[c(2,10,20)])
+col_fun_CIN <- colorRamp2(c(10, 13, 15), c(""white"", '#DBBEBE', '#B75C5B'))
+col_fun_sTILS <- colorRamp2(c(0, 0.1, 0.35), c(""white"", '#FDF2C9', '#F8CA26'))
+
+## heatmap
+p1 = Heatmap(mat1[,ord],
+        cluster_rows = FALSE,cluster_columns = FALSE,
+        col = col_fun,
+        na_col = ""#eaeaea"",
+        border = T,
+        show_column_names = FALSE,
+        column_split =factor(luminal$SNF, levels = paste0(""SNF"",1:4)) ,
+        column_gap = unit(2, ""mm""),
+        row_names_side = ""left"",
+        heatmap_legend_param = list(title = ""Z-score""),
+        top_annotation = HeatmapAnnotation(SNF = luminal$SNF,
+                                           PAM50 = luminal$PAM50_classifier,
+                                           CINscore = luminal$CINscore,
+                                           sTils = luminal$sTILs,
+                                           na_col = ""#eaeaea"",
+                                           gap = unit(1, ""points""),
+                                           annotation_name_side = ""left"",
+                                           annotation_name_gp = gpar(fontsize = 9),
+                                           col = list(PAM50 = color_PAM50,
+                                                      SNF = color,
+                                                      CINscore = col_fun_CIN,
+                                                      sTils = col_fun_sTILS
+                                                      #Relapse = c('0' = 'white', '1' = 'black')
+                                                      ))
+        )
+p2 = Heatmap(mat2[,ord],
+             cluster_rows = FALSE,cluster_columns = FALSE,
+             col = col_fun,
+             na_col = ""#eaeaea"",
+             border = T,
+             row_names_side = ""left"",
+             show_heatmap_legend = FALSE,
+             show_column_names = FALSE,
+             column_split =factor(luminal$SNF, levels = paste0(""SNF"",1:4)) ,
+             column_gap = unit(2, ""mm""))
+
+p3 = Heatmap(mat3[,ord],
+             cluster_rows = FALSE,cluster_columns = FALSE,
+             col = col_fun,
+             border = T,
+             na_col = ""#eaeaea"",
+             row_names_side = ""left"",
+             show_heatmap_legend = FALSE,
+             show_column_names = FALSE,
+             column_split =factor(luminal$SNF, levels = paste0(""SNF"",1:4)) ,
+             column_gap = unit(2, ""mm""))
+
+p4 = Heatmap(mat4[,ord],
+             na_col = ""#eaeaea"",
+             cluster_rows = FALSE,cluster_columns = FALSE,
+             col = col_fun,
+             border = T,
+             row_names_side = ""left"",
+             show_heatmap_legend = FALSE,
+             show_column_names = FALSE,
+             column_split =factor(luminal$SNF, levels = paste0(""SNF"",1:4)) ,
+             column_gap = unit(2, ""mm""))
+
+p5 = Heatmap(mat5[,ord],
+             na_col = ""#eaeaea"",
+             cluster_rows = FALSE,cluster_columns = FALSE,
+             col = col_fun,
+             border = T,
+             row_names_side = ""left"",
+             show_heatmap_legend = FALSE,
+             show_column_names = FALSE,
+             column_split =factor(luminal$SNF, levels = paste0(""SNF"",1:4)) ,
+             column_gap = unit(2, ""mm""))
+
+
+p = p1 %v% p2  %v% p3 %v% p4 %v% p5
+p
+
+export::graph2pdf(p,file = ""Fig5.pdf"",width = 12,height = 8)
+export::graph2ppt(p,file = ""Fig5.pptx"",width = 12,height = 8)
+
+
+### anova
+mat_merge = rbind(as.data.frame(t(luminal[ord,c(""CINscore"",""sTILs"")])),mat1[,ord], mat2[,ord], mat3[,ord], mat4[,ord], mat5[,ord])
+t.test.res = as.data.frame(matrix(nrow = nrow(mat_merge),ncol = 2))
+rownames(t.test.res) = rownames(mat_merge) ; colnames(t.test.res) = c(""p.val"",'adj.p')
+for ( i in rownames(mat_merge)){
+  tmp0 = luminal$SNF ; names(tmp0) = row.names(luminal)
+  tmp = data.frame(gene = as.numeric(mat_merge[i,rownames(luminal)]),
+                   group = luminal$SNF)
+  res = summary(aov(gene~group,data = tmp))
+  t.test.res[i,""p.val""] = res[[1]][[""Pr(>F)""]][1]
+}
+t.test.res$adj.p = p.adjust(t.test.res$p.val,method = ""fdr"")
+
+write.csv(t.test.res,file = ""aov_res.csv"")
+
+","R"
+"Receptor","yifanzhou330/Luminal-SNF","Fig5_and_related_ED_Fig/Integrated_immune_landscape_external_validation.R",".R","1844","49","rm(list = ls()) ; graphics.off()
+library(RColorBrewer)
+library(pheatmap)
+library(export)
+library(ggpubr)
+
+load(""TCGA_hg38_FPKM_Mut_hg19ascatCNV.Rdata"")
+SNF.file = ""RawNum100_CorThresh0.75_RF_scale_fold0_TCGA""
+SNFnew = openxlsx::read.xlsx(paste0(SNF.file,"".xlsx""),1, rowNames = T)
+SNF_Cluster = SNFnew$inferSNF ; names(SNF_Cluster) = rownames(SNFnew)
+
+exp.fpkm.TT.log = log2(exp_fpkm_TCGA +1)
+
+# PDCD1
+PDCD1 = as.data.frame(t(exp.fpkm.TT.log[""PDCD1"",names(SNF_Cluster)]))
+PDCD1$SNF = SNF_Cluster
+PDCD1$SNF[PDCD1$SNF != ""SNF2""] = ""Others""
+PDCD1$SNF = factor(PDCD1$SNF,levels = c(""SNF2"",""Others""))
+
+plot1 = ggviolin(data = PDCD1,x = ""SNF"", y = ""PDCD1"",fill = ""SNF"",
+         add = ""boxplot"", add.params = list(fill=""white""),
+          palette = c(color[2], ""grey"") ) + stat_compare_means(label.y = 4)
+graph2pdf(plot1, file = ""PDCD1_SNF2_vs_Others.pdf"",height = 4,width = 4)  
+
+
+## hp
+immune  = c(""CD8A"",""GZMA"",""PRF1"",""IDO1"")
+
+tmp = c(""#2378B3"",  ""#1CB038"",  ""#F8A900"", ""#D5271A"")
+names(tmp) = paste0(""SNF"",1:4)
+annotation_color = c(list(SNF_Cluster =tmp))
+
+annot_col = as.data.frame(SNF_Cluster)
+
+immune.hp = exp_fpkm_TCGA[immune,names(SNF_Cluster)]
+bk = unique(c(seq(-1,1,length = 100)))
+p = pheatmap::pheatmap(immune.hp[,names(sort(SNF_Cluster))],cluster_rows = F,cluster_cols = F,
+                       scale = ""row"", show_colnames = F,
+                       annotation_col =annot_col, annotation_colors = annotation_color,
+                       border_color = NA,
+                       breaks = bk,
+                       gaps_col = cumsum(as.matrix(table(SNF_Cluster))[,1])[-4],
+                       gaps_row = 1:3,
+                       cellheight = 15,cellwidth = 3,
+                       filename = ""FPKM_SNF_immune_hp.pdf"",
+                       height = 10)
+graph2ppt(p,file = ""FPKM_SNF_immune_hp.ppt"",height = 10, width = 25)
+
+","R"
+"Receptor","yifanzhou330/Luminal-SNF","Fig4_and_related_ED_Fig/Integrated_CellCycle_landscape.R",".R","8427","185","rm(list = ls())
+graphics.off()
+library(maftools)
+library(ComplexHeatmap)
+library(circlize)
+library(stringr)
+library(RColorBrewer)
+load(""CBCGA_HRposHER2neg354_WES_RNAseq_CNV_Metab_Protein_20211127.Rdata"")
+exp.fpkm.TT.log = log2(exp.fpkm.TT+1)
+#### add infor  -------
+MGPS = read.csv(""MGPS_gene.csv"")[,2]
+MGPS = exp.fpkm.TT[intersect(MGPS,rownames(exp.fpkm.TT)),]
+MGPS = apply(MGPS,2,mean)
+
+CINscore = as.data.frame(matrix(nrow = length(SNF_Cluster), ncol = 2))
+rownames(CINscore) = names(sort(SNF_Cluster))  ;colnames(CINscore) = c(""CINscore"",""SNF_All"")
+CINscore$SNF_All = paste0(""SNF"",sort(SNF_Cluster))
+for (i in names(sort(SNF_Cluster))){
+  tmp = cnv.alldata[,i]
+  tmp = tmp^2
+  CINscore[i,""CINscore""] = sum(tmp)
+}
+CINscore$CINscore = log2(CINscore$CINscore)
+
+#### annot bar -------
+luminal$PAM50_PC_TCGA = factor(luminal$PAM50_PC_TCGA, levels =c( ""LumA"" ,""LumB"" ,""Her2"",""Basal"" ,""Normal""))
+luminal$HRD = as.numeric(luminal$HRD)
+luminal$MGPS = MGPS[rownames(luminal)]
+luminal$MGPS.scale = scale(luminal$MGPS)
+luminal$SNF3orNot = luminal$SNF分型
+luminal$SNF3orNot[luminal$SNF3orNot != ""SNF3""] = ""non_SNF3""
+luminal$CINscore = CINscore[rownames(luminal),""CINscore""]
+
+protein_log2_NA = as.data.frame(matrix(NA,nrow = nrow(protein_log2), 
+                                       ncol = length(setdiff(rownames(luminal),colnames(protein_log2))) ))
+rownames(protein_log2_NA) = rownames(protein_log2)
+colnames(protein_log2_NA) = setdiff(rownames(luminal),colnames(protein_log2))
+
+
+#### mat1 ---------
+tmp = c(""CCND1"",""CDK2"") #tmp = c(""CCND1"",""CDK1"",""CDK2"")
+mat1_RNA = exp.fpkm.TT.log[tmp, names(sort(SNF_Cluster))]
+mat1_RNA = t(scale(t(mat1_RNA)))
+rownames(mat1_RNA) = paste0(rownames(mat1_RNA),""_RNA"")
+
+# mat1_CNV = cnv.thre[tmp, names(sort(SNF_Cluster))]
+# mat1_CNV = apply(mat1_CNV, 2, function(x){ifelse(x < 1, 0, 2)} )
+mat1_CNV = cnv.alldata[tmp, names(sort(SNF_Cluster))]
+mat1_CNV = apply(mat1_CNV, 2, function(x){ifelse(x < log2(4/2), 0, 2)} )
+rownames(mat1_CNV) = paste0(rownames(mat1_CNV),""_CNA"")
+
+mat1_Pro = protein_log2[intersect(tmp,rownames(protein_log2)), ]
+mat1_Pro = t(scale(t(mat1_Pro)))
+mat1_Pro = cbind(mat1_Pro,protein_log2_NA[rownames(mat1_Pro),])
+mat1_Pro = mat1_Pro[,names(sort(SNF_Cluster))]
+rownames(mat1_Pro) = paste0(rownames(mat1_Pro),""_Protein"")
+
+mat1 = rbind(mat1_CNV,mat1_RNA,mat1_Pro)
+mat1 = mat1[c(""CCND1_CNA"",""CCND1_RNA"",""CCND1_Protein"",
+#              ""CDK1_CNA"",""CDK1_RNA"",""CDK1_Protein"",
+              ""CDK2_CNA"",""CDK2_RNA"",""CDK2_Protein""),]
+
+#### mat2 ---------
+mat2_RNA = exp.fpkm.TT.log[c(""CDK1""), names(sort(SNF_Cluster))]
+mat2_RNA = t(scale(t(mat2_RNA)))
+rownames(mat2_RNA) = paste0(rownames(mat2_RNA),""_RNA"")
+
+mat2_CNV = cnv.alldata[c(""CDK1""), names(sort(SNF_Cluster))]
+mat2_CNV = t(as.data.frame(apply(mat2_CNV, 2, function(x){ifelse(x < log2(4/2), 0, 2)} )))
+rownames(mat2_CNV) = ""CDK1_CNA""
+
+mat2_Pro = protein_log2[c(""CDK1""), ]
+mat2_Pro = t(scale(t(mat2_Pro)))
+mat2_Pro = cbind(mat2_Pro,protein_log2_NA[c(""CDK1""),])
+mat2_Pro = mat2_Pro[,names(sort(SNF_Cluster))]
+rownames(mat2_Pro) = paste0(rownames(mat2_Pro),""_Protein"")
+
+mat2 = rbind(mat2_RNA,mat2_CNV,mat2_Pro)
+
+
+#### mat3 ---------
+tmp = c(""AURKA"",""MYBL2"",""TOP2A"",""ESPL1"",""DSCC1"",""GINS4"",""RAD21"",""KIF18B"",""E2F1"",""E2F2"")
+mat3_RNA = exp.fpkm.TT.log[tmp, names(sort(SNF_Cluster))]
+mat3_RNA = t(scale(t(mat3_RNA)))
+rownames(mat3_RNA) = paste0(rownames(mat3_RNA),""_RNA"")
+
+mat3_Pro = protein_log2[intersect(rownames(protein_log2),tmp), ]
+mat3_Pro = t(scale(t(mat3_Pro)))
+mat3_Pro = cbind(mat3_Pro,protein_log2_NA[rownames(mat3_Pro),])
+mat3_Pro = mat3_Pro[,names(sort(SNF_Cluster))]
+rownames(mat3_Pro) = paste0(rownames(mat3_Pro),""_Protein"")
+
+mat3 = rbind(mat3_RNA,mat3_Pro)
+mat3 = mat3[sort(rownames(mat3)),]
+mat3 = mat3[c(2,1,4,3,5,6,7,9,8,10,11,13,12,15,14) ,]
+
+#### draw  ------
+## ord
+ord = rownames(luminal[order(luminal$SNF3orNot,luminal$MGPS.scale,decreasing = F),])
+luminal = luminal[ord,]
+
+## annot 
+color_PAM50 = c(""#1D76BC"",""#76CFE6"",""#6E59A6"",""#E11D2E"",""#CDCFD0"")
+names(color_PAM50) = c(""LumA"",""LumB"",""Her2"",""Basal"",""Normal"")
+names(color) = paste0(""SNF"",1:4)
+col_fun = colorRamp2(c(-2,-1,0 , 1,2), rev(brewer.pal(n = 5, name =""RdBu"")))
+#lgd = Legend(col_fun = col_fun,title = ""z-score"")
+# col_fun_ki67 <- colorRamp2(c(3, 30, 98), c('#EBEFF6', '#99AFD2', '#3C74AE'))
+col_fun_CIN <- colorRamp2(c(10, 13, 15), c(""white"", '#DBBEBE', '#B75C5B'))
+col_fun_MGPS <-  colorRamp2(c(-1.5,-0.5, 0 , 0.5,1.5), rev(brewer.pal(n = 5, name =""RdBu"")))
+col_fun_Isig <- colorRamp2(c(1, 1.1 ,1.2), colorRampPalette(c(""white"", ""#08506F""))(20)[c(2,10,20)])
+col_fun_SSig <- colorRamp2(c(1, 1.1 ,1.2), colorRampPalette(c(""white"", ""#50A9AC""))(20)[c(2,10,20)])
+col_fun_hrd <- colorRamp2(c(0, 16, 75), c('#FDF9E7', '#EFE089', '#D7C801'))
+
+## heatmap
+p1 = Heatmap(mat1[,ord],
+             cluster_rows = FALSE,cluster_columns = FALSE,
+             col = col_fun,
+             na_col = ""#eaeaea"",
+             border = T,
+             show_column_names = FALSE,
+             column_split =factor(luminal$SNF3orNot, levels = c(""SNF3"",""non_SNF3"")) ,
+             column_gap = unit(2, ""mm""),
+             row_names_side = ""left"",
+             heatmap_legend_param = list(title = ""Z-score""),
+             top_annotation = HeatmapAnnotation(SNF = luminal$SNF分型,
+                                                PAM50 = luminal$PAM50_PC_TCGA,
+                                                #ImmuneSignature = luminal$ImmuneSignature,
+                                                #StromalSignature = luminal$StromalSignature,
+                                                CINscore = luminal$CINscore,
+                                                HRDscore = luminal$HRD,
+                                                MGPS = luminal$MGPS.scale,
+                                                # Age.at.surgery = CBCGAClin$Age,
+                                                # Lymph_node_status = CBCGAClin$Lymph_node_status,
+                                                # Ki67 = as.numeric(CBCGAClin$Ki67),
+                                                #Relapse = luminal$DMFS_status ,
+                                                #border = T,
+                                                gap = unit(1, ""points""),
+                                                annotation_name_side = ""left"",
+                                                annotation_name_gp = gpar(fontsize = 9),
+                                                col = list(PAM50 = color_PAM50,
+                                                           SNF = color,
+                                                           #StromalSignature = col_fun_SSig,
+                                                           #ImmuneSignature =col_fun_Isig,
+                                                           CINscore = col_fun_CIN,
+                                                           #Age.at.surgery = col_fun_age,
+                                                           #Lymph_node_status = c('0' = 'white', '1' = 'black'),
+                                                           #Ki67 = col_fun_ki67,
+                                                           HRDscore = col_fun_hrd,
+                                                           MGPS = col_fun_MGPS
+                                                           #Relapse = c('0' = 'white', '1' = 'black')
+                                                ))
+)
+p2 = Heatmap(mat2[,ord],
+             cluster_rows = FALSE,cluster_columns = FALSE,
+             col = col_fun,
+             na_col = ""#eaeaea"",
+             border = T,
+             row_names_side = ""left"",
+             show_heatmap_legend = FALSE,
+             show_column_names = FALSE,
+             column_split = factor(luminal$SNF分型, levels = c('SNF1', 'SNF2', 'SNF3', 'SNF4')),
+             column_gap = unit(2, ""mm""))
+
+p3 = Heatmap(mat3[,ord],
+             cluster_rows = FALSE,cluster_columns = FALSE,
+             col = col_fun,
+             na_col = ""#eaeaea"",
+             border = T,
+             row_names_side = ""left"",
+             show_heatmap_legend = FALSE,
+             show_column_names = FALSE,
+             column_split = factor(luminal$SNF分型, levels = c('SNF1', 'SNF2', 'SNF3', 'SNF4')),
+             column_gap = unit(2, ""mm""))
+
+p = p1 %v% p2 %v% p3
+p
+
+export::graph2pdf(p,file = ""Fig4_alldata4_V2.pdf"",width = 12,height = 7)
+export::graph2ppt(p,file = ""Fig4_alldata4_V2.pptx"",width = 12,height = 7)
+
+
+
+
+","R"
+"Receptor","yifanzhou330/Luminal-SNF","Fig4_and_related_ED_Fig/Integrated_CellCycle_landscape_external_validation.R",".R","7828","206","rm(list = ls()) ; graphics.off()
+
+library(RColorBrewer)
+library(pheatmap)
+library(export)
+CDK = c(""CCND1"",""CDK1"",""CDK2"")
+load(""/TCGA_hg38_FPKM_Mut_hg19ascatCNV.Rdata"")
+SNF.file = ""RawNum100_CorThresh0.75_RF_scale_fold0_TCGA""
+SNFnew = openxlsx::read.xlsx(paste0(SNF.file,"".xlsx""),1,rowNames = T)
+SNF_Cluster = SNFnew$inferSNF ; names(SNF_Cluster) = rownames(SNFnew)
+
+#### RNA -------- 
+CDK.hp = exp_fpkm_TCGA[CDK,names(SNF_Cluster)]
+bk = unique(c(seq(-1,1,length = 100)))
+p = pheatmap::pheatmap(CDK.hp[,names(sort(SNF_Cluster))],cluster_rows = F,cluster_cols = F,
+                       scale = ""row"", show_colnames = F,
+                       #annotation_col =annot_col, annotation_colors = annotation_color,
+                       border_color = NA,
+                       breaks = bk,
+                       gaps_col = cumsum(as.matrix(table(SNF_Cluster))[,1])[-4],
+                       gaps_row = 1:2,
+                       cellheight = 15,cellwidth = 3,
+                       filename = ""FPKM_TCGAinferSNF_CDK_hp.pdf"",
+                       height = 10)
+graph2ppt(p,file = ""FPKM_TCGAinferSNF_CDK_hp.ppt"",height = 10, width = 20)
+
+## stat
+CDK.hp = as.data.frame(t(CDK.hp))
+CDK.hp$SNF = paste0(""SNF"",SNF_Cluster )
+
+aov.res = as.data.frame(matrix(nrow = length(CDK), ncol = 2))
+rownames(aov.res) = CDK ; colnames(aov.res) = c(""p.val"",""adj.p"")
+
+for (i in CDK){
+  mat = CDK.hp[,c(i,""SNF"")]
+  colnames(mat)[1] = ""gene""
+  shapiro.test(mat[,""gene""])
+  bartlett.test(gene~SNF,data=mat)
+  res = summary(aov(gene~SNF,data=mat))
+  aov.res[i,""p.val""] = res[[1]][[""Pr(>F)""]][1]
+}
+aov.res[,""adj.p""] = p.adjust(aov.res[,""p.val""],method = ""fdr"")
+write.csv(aov.res, file =  ""FPKM_TCGAinferSNF_CDK_hp_aov.csv"")
+
+
+#### CNV ------
+CDK.hp = cnv_thre_hg19ascat_TCGA[CDK,names(SNF_Cluster)]
+for (i in rownames(CDK.hp)){
+  tmp = CDK.hp[i,]
+  tmp[tmp < 0 ] = 0
+  tmp[tmp > 0 ] = 1
+  CDK.hp[i,] = tmp
+}
+annot_col = data.frame(row.names = names(SNF_Cluster), subtype =SNF_Cluster )
+tmp = c(""#2378B3"",  ""#1CB038"",  ""#F8A900"", ""#D5271A"")
+names(tmp) = c(paste0(""SNF"",1:4))
+annotation_color = c(list(subtype =tmp))
+
+bk = unique(c(seq(-1,1,length = 50)))
+p = pheatmap::pheatmap(CDK.hp[,names(sort(SNF_Cluster))],cluster_rows = F,cluster_cols = F,
+                       show_colnames = F,
+                       annotation_col =annot_col,
+                       border_color = NA,
+                       breaks = bk,
+                       gaps_col = cumsum(as.matrix(table(SNF_Cluster))[,1])[-4],
+                       gaps_row = 1:2,
+                       color =  colorRampPalette(rev(brewer.pal(n = 5, name = ""RdBu"")))(50),
+                       annotation_colors = annotation_color,
+                       cellheight = 15,cellwidth = 3,
+                       filename = ""CNV_SNF_CDK_hp.pdf"",
+                       height = 10)
+graph2ppt(p,file = ""CNV_SNF_CDK_hp.ppt"",height = 10, width = 20)
+
+## stat
+CDK.hp = as.data.frame(t(CDK.hp))
+CDK.hp$SNF = paste0(""SNF"",SNF_Cluster )
+
+fisher.res = as.data.frame(matrix(nrow = length(CDK), ncol = 2))
+rownames(fisher.res) = CDK ; colnames(fisher.res) = c(""p.val"",""adj.p"")
+
+for (i in CDK){
+  mat = CDK.hp[,c(i,""SNF"")]
+  colnames(mat)[1] = ""gene""
+  mytab = xtabs( ~ gene+SNF,data=mat)
+  res = fisher.test( mytab)
+  fisher.res[i,""p.val""] = res[[""p.value""]]
+}
+fisher.res[,""adj.p""] = p.adjust(fisher.res[,""p.val""],method = ""fdr"")
+write.csv(fisher.res, file =  ""CNV_TCGAinferSNF_CDK_fisher.csv"")
+
+
+
+# MDM2 RNA -------
+i= ""MDM2""
+exp.fpkm.TT.log = log2(exp_fpkm_TCGA+1)
+gene = as.data.frame(t(exp.fpkm.TT.log[i,]))
+colnames(gene ) = ""gene""
+gene$SNF = SNF_Cluster[colnames(exp.fpkm.TT.log)]
+gene$SNF[gene$SNF != ""SNF3""] = ""Others""
+gene$SNF = factor(gene$SNF, levels = c(""SNF3"",""Others""))
+
+shapiro.test(gene$gene)
+bartlett.test(gene~SNF,data=gene)
+
+p = ggboxplot(gene,x = ""SNF"",y= ""gene"",color = ""SNF"",
+              add = ""jitter"", 
+              palette = c(color[3], ""grey"") )+
+  stat_compare_means(label.y = 3.5) 
+ggsave(p, filename = paste0(""./SNF3_"",i,""_FPKMlog.pdf""),height = 4,width = 4)
+graph2ppt(p,file = paste0(""./SNF3_"",i,""_FPKMlog.ppt""),height = 4,width = 4)
+
+# ATM RNA --------
+i= ""ATM""
+exp.fpkm.TT.log = log2(exp_fpkm_TCGA+1)
+gene = as.data.frame(t(exp.fpkm.TT.log[i,]))
+colnames(gene ) = ""gene""
+gene$SNF = SNF_Cluster[colnames(exp.fpkm.TT.log)]
+gene$SNF[gene$SNF != ""SNF3""] = ""Others""
+gene$SNF = factor(gene$SNF, levels = c(""SNF3"",""Others""))
+
+shapiro.test(gene$gene)
+bartlett.test(gene~SNF,data=gene)
+
+p = ggboxplot(gene,x = ""SNF"",y= ""gene"",color = ""SNF"",
+              add = ""jitter"", 
+              palette = c(color[3], ""grey"") )+
+  stat_compare_means(label.y = 3) 
+ggsave(p, filename = paste0(""./SNF3_"",i,""_FPKMlog.pdf""),height = 4,width = 4)
+graph2ppt(p,file = paste0(""./SNF3_"",i,""_FPKMlog.ppt""),height = 4,width = 4)
+
+
+# MDM2 AMP -----
+i = ""MDM2""
+cnv = as.data.frame(t(cnv_thre_hg19ascat_TCGA[i,names(SNF_Cluster)]))
+colnames(cnv) = ""gene""
+cnv$gene[cnv$gene < 0] = 0
+cnv$gene[cnv$gene > 1 ] = 1
+cnv$SNF = SNF_Cluster
+cnv$SNF[cnv$SNF != ""SNF3""] = ""Others""
+cnv$SNF = factor(cnv$SNF, levels = c(""SNF3"",""Others""))
+
+cluster.merge4plot = as.data.frame(table(cnv$SNF, cnv$gene))
+colnames(cluster.merge4plot)[1:2] = c(""SNF"",""gene"")
+
+cluster.merge4plot2 = ddply(cluster.merge4plot,""SNF"", transform,
+                            percent = Freq / sum(Freq) *100)
+cluster.merge4plot2$gene = factor(cluster.merge4plot2$gene,levels = c(0, 1)) 
+cnv_color = c( ""1"" = ""#E21F22"",""0"" = ""#eaeaea"")
+
+tmp = fisher.test(table(cnv$SNF, cnv$gene))
+p = ggplot(cluster.merge4plot2, aes(x=SNF, y = percent, fill = gene )) + geom_bar(stat = ""identity"") +
+  scale_fill_manual(values = cnv_color)+
+  theme(panel.border = element_rect(fill=NA, colour = ""black"", size=2), 
+        axis.ticks = element_line(size = 2),
+        axis.text.x=element_text(size = 20),
+        axis.text.y=element_text(size = 20),
+        axis.title.x = element_text(size = 20),
+        axis.title.y = element_text(size = 20),
+        panel.background = element_blank(),
+        panel.grid.major = element_blank(), 
+        panel.grid.minor = element_blank() 
+  ) +
+  ggtitle(paste0(""p = "",round(tmp[[""p.value""]],5)) ) 
+ggsave(p,filename = paste0(""./SNF3_"",i,""_4plot_percent_barplot.pdf""))
+graph2ppt(p, file = paste0(""./SNF3_"",i,""_4plot_percent_barplot.ppt""))
+
+
+# ATM Del -----
+i = ""ATM""
+cnv = as.data.frame(t(cnv_thre_hg19ascat_TCGA[i,names(SNF_Cluster)]))
+colnames(cnv) = ""gene""
+cnv$gene[cnv$gene > 0] = 0
+cnv$gene[cnv$gene < 0 ] = 1
+cnv$SNF = SNF_Cluster
+cnv$SNF[cnv$SNF != ""SNF3""] = ""Others""
+cnv$SNF = factor(cnv$SNF, levels = c(""SNF3"",""Others""))
+
+cluster.merge4plot = as.data.frame(table(cnv$SNF, cnv$gene))
+colnames(cluster.merge4plot)[1:2] = c(""SNF"",""gene"")
+
+cluster.merge4plot2 = ddply(cluster.merge4plot,""SNF"", transform,
+                            percent = Freq / sum(Freq) *100)
+cluster.merge4plot2$gene = factor(cluster.merge4plot2$gene,levels = c(0, 1)) 
+#cnv_color = c(""-2"" = ""#3654A4"", ""2"" = ""#E21F22"",""-1"" = ""#BBC3D9"",""1"" =""#E4B5B8"" ,""0"" = ""#eaeaea"")
+cnv_color = c( ""1"" = ""#3654A4"",""0"" = ""#eaeaea"")
+
+tmp = fisher.test(table(cnv$SNF, cnv$gene))
+p = ggplot(cluster.merge4plot2, aes(x=SNF, y = percent, fill = gene )) + geom_bar(stat = ""identity"") +
+  scale_fill_manual(values = cnv_color)+
+  theme(panel.border = element_rect(fill=NA, colour = ""black"", size=2), 
+        axis.ticks = element_line(size = 2),
+        axis.text.x=element_text(size = 20),
+        axis.text.y=element_text(size = 20),
+        axis.title.x = element_text(size = 20),
+        axis.title.y = element_text(size = 20),
+        panel.background = element_blank(),
+        panel.grid.major = element_blank(), 
+        panel.grid.minor = element_blank() 
+  ) +
+  ggtitle(paste0(""p = "",round(tmp[[""p.value""]],5)) ) 
+ggsave(p,filename =  paste0(""./SNF3_"",i,""_4plot_percent_barplot.pdf""))
+graph2ppt(p, file = paste0(""./SNF3_"",i,""_4plot_percent_barplot.ppt""))
+
+
+
+","R"
+"Receptor","yifanzhou330/Luminal-SNF","Fig1_and_related_ED_Fig/fdr_hp.R",".R","4964","113","rm(list = ls()) ; graphics.off()
+load(""CBCGA_HRposHER2neg351_WES_RNAseq_CNV_Metab_Protein_20220829.Rdata"")
+library(pheatmap) ; library(RColorBrewer)
+##### fdr val for pheatmap validation---
+## fdr_RNA_NMF_351 
+tmp1 = read.csv(""multiomic_feature.pdf_clinic_aov.csv"",
+                row.names = 1)
+tmp2 = read.csv(""multiomic_feature.pdf_mut_chisq.csv"",
+                row.names = 1)
+tmp3 = read.csv(""multiomic_feature.pdf_AMP_chisq.csv"",
+                row.names = 1)
+tmp4 = read.csv(""multiomic_feature.pdf_path_aov.csv"",
+                row.names = 1)
+tmp5 = read.csv(""multiomic_feature.pdf_metab_aov.csv"",
+                row.names = 1)
+
+fdr_RNA_NMF_351 = -log10(c(tmp1$adj.p, tmp2$adj.p, tmp3$adj.p, tmp4$adj.p,tmp5$adj.p))
+names(fdr_RNA_NMF_351) = c(rownames(tmp1) ,rownames(tmp2),rownames(tmp3),rownames(tmp4),rownames(tmp5))
+fdr_RNA_NMF_351[fdr_RNA_NMF_351 >= 10] = 10
+fdr_RNA_NMF_351[fdr_RNA_NMF_351 <= -log10(0.05) ] = NA
+
+## fdr_CNA_RNA
+tmp1 = read.csv(""multiomic_feature.pdf_clinic_aov.csv"",
+                row.names = 1)
+tmp2 = read.csv(""multiomic_feature.pdf_mut_chisq.csv"",
+                row.names = 1)
+tmp3 = read.csv(""multiomic_feature.pdf_AMP_chisq.csv"",
+                row.names = 1)
+tmp4 = read.csv(""multiomic_feature.pdf_path_aov.csv"",
+                row.names = 1)
+tmp5 = read.csv(""multiomic_feature.pdf_metab_aov.csv"",
+                row.names = 1)
+
+fdr_RNA_CNA = -log10(c(tmp1$adj.p, tmp2$adj.p, tmp3$adj.p, tmp4$adj.p,tmp5$adj.p))
+names(fdr_RNA_CNA) = c(rownames(tmp1) ,rownames(tmp2),rownames(tmp3),rownames(tmp4),rownames(tmp5))
+fdr_RNA_CNA[fdr_RNA_CNA >= 10] = 10
+fdr_RNA_CNA[fdr_RNA_CNA <= -log10(0.05) ] = NA
+
+
+## fdr WES_RNA_CNA_Polar
+tmp1 = read.csv(""multiomic_feature.pdf_clinic_aov.csv"",
+                row.names = 1)
+tmp2 = read.csv(""multiomic_feature.pdf_mut_chisq.csv"",
+                row.names = 1)
+tmp3 = read.csv(""multiomic_feature.pdf_AMP_chisq.csv"",
+                row.names = 1)
+tmp4 = read.csv(""multiomic_feature.pdf_path_aov.csv"",
+                row.names = 1)
+tmp5 = read.csv(""multiomic_feature.pdf_metab_aov.csv"",
+                row.names = 1)
+
+fdr_WES_RNA_CNA_Polar = -log10(c(tmp1$adj.p, tmp2$adj.p, tmp3$adj.p, tmp4$adj.p,tmp5$adj.p))
+names(fdr_WES_RNA_CNA_Polar) = c(rownames(tmp1) ,rownames(tmp2),rownames(tmp3),rownames(tmp4),rownames(tmp5))
+fdr_WES_RNA_CNA_Polar[fdr_WES_RNA_CNA_Polar >= 10] = 10
+fdr_WES_RNA_CNA_Polar[fdr_WES_RNA_CNA_Polar <= -log10(0.05) ] = NA
+
+
+## fdr_oriSNF
+tmp1 = read.csv(""Fig2_multiomic_feature.pdf_clinic_aov.csv"", row.names = 1)
+tmp2 = read.csv(""Fig2_multiomic_feature.pdf_mut_chisq.csv"", row.names = 1)
+tmp3 = read.csv(""Fig2_multiomic_feature.pdf_AMP_chisq.csv"", row.names = 1)
+tmp4 = read.csv(""Fig2_multiomic_feature.pdf_path_aov.csv"", row.names = 1)
+tmp5 = read.csv(""Fig2_multiomic_feature.pdf_metab_aov.csv"", row.names = 1)
+
+
+fdr_oriSNF = -log10(c(tmp1$adj.p, tmp2$adj.p, tmp3$adj.p, tmp4$adj.p,tmp5$adj.p))
+names(fdr_oriSNF) = c(rownames(tmp1) ,rownames(tmp2),rownames(tmp3),rownames(tmp4),rownames(tmp5))
+fdr_oriSNF[fdr_oriSNF >= 10] = 10
+fdr_oriSNF[fdr_oriSNF <= -log10(0.05) ] = NA
+
+
+
+## fdr Pro_RNA_CNA_Polar
+tmp1 = read.csv(""multiomic_feature.pdf_clinic_aov.csv"",
+                row.names = 1)
+tmp2 = read.csv(""multiomic_feature.pdf_mut_chisq.csv"",
+                row.names = 1)
+tmp3 = read.csv(""multiomic_feature.pdf_AMP_chisq.csv"",
+                row.names = 1)
+tmp4 = read.csv(""multiomic_feature.pdf_path_aov.csv"",
+                row.names = 1)
+tmp5 = read.csv(""multiomic_feature.pdf_metab_aov.csv"",
+                row.names = 1)
+
+fdr_Pro_RNA_CNA_Polar = -log10(c(tmp1$adj.p, tmp2$adj.p, tmp3$adj.p, tmp4$adj.p,tmp5$adj.p))
+names(fdr_Pro_RNA_CNA_Polar) = c(rownames(tmp1) ,rownames(tmp2),rownames(tmp3),rownames(tmp4),rownames(tmp5))
+fdr_Pro_RNA_CNA_Polar[fdr_Pro_RNA_CNA_Polar >= 10] = 10
+fdr_Pro_RNA_CNA_Polar[fdr_Pro_RNA_CNA_Polar <= -log10(0.05) ] = NA
+
+
+##
+fdr_Pro_RNA_CNA_Polar = fdr_Pro_RNA_CNA_Polar[intersect(names(fdr_Pro_RNA_CNA_Polar),names(fdr_oriSNF))]
+
+fdr_RNA_NMF_351 = as.data.frame(fdr_RNA_NMF_351[names(fdr_Pro_RNA_CNA_Polar)])
+fdr_RNA_CNA = as.data.frame(fdr_RNA_CNA[names(fdr_Pro_RNA_CNA_Polar)])
+fdr_oriSNF = as.data.frame(fdr_oriSNF[names(fdr_Pro_RNA_CNA_Polar)])
+fdr_WES_RNA_CNA_Polar = as.data.frame(fdr_WES_RNA_CNA_Polar[names(fdr_Pro_RNA_CNA_Polar)])
+fdr_Pro_RNA_CNA_Polar = as.data.frame(fdr_Pro_RNA_CNA_Polar)
+
+fdr = data.frame(row.names = row.names(fdr_RNA_NMF_351),
+                 RNA_NMF_351 = fdr_RNA_NMF_351[,1],
+                 RNA_CNA = fdr_RNA_CNA[,1],
+                 oriSNF = fdr_oriSNF[,1],
+                 WES_RNA_CNA_Polar = fdr_WES_RNA_CNA_Polar[,1],
+                 Pro_RNA_CNA_Polar = fdr_Pro_RNA_CNA_Polar[,1])
+pheatmap::pheatmap(fdr,cluster_rows = F,cluster_cols = F,display_numbers = T,
+                   colorRampPalette(brewer.pal(n = 7, name = ""Greys"")[1:5])(100), #YlOrRd
+                   na_col = ""#eaeaea"",
+                   border  = NA,filename = ""./fdr_hp.pdf"",height = 8,width = 7.8)
+
+
+
+","R"
+"Receptor","yifanzhou330/Luminal-SNF","Fig1_and_related_ED_Fig/Define_SNF_cluster.R",".R","3016","107","rm(list = ls() ) ; graphics.off()
+library(paletteer) 
+library(export)
+library(ggpubr)
+load(""./CBCGA_HRposHER2neg351_WES_RNAseq_CNV_Metab_Protein_20221010.Rdata"")
+################################
+# data preapre 
+################################
+# cnv
+cnv = cnv.alldata
+
+# exp
+exp.fpkm.TT.log = log2(exp.fpkm.TT +1)
+exp.fpkm.TT.log.SD2k = exp.fpkm.TT.log[order(apply(exp.fpkm.TT.log,1,sd),decreasing = T)[1:2000],] 
+
+# polar
+polar = polar_metabolite_TT_MS2_log2
+
+# data transform
+cnv = as.data.frame(t(cnv))
+exp.fpkm.TT.log.SD2k = as.data.frame(t(exp.fpkm.TT.log.SD2k))
+polar = as.data.frame(t(polar))
+
+all(rownames(cnv) == rownames(exp.fpkm.TT.log.SD2k)) 
+all(rownames(cnv) == rownames(polar)) 
+
+rm(list = ls()[! ls() %in% c(""luminal"",""cnv"",""exp.fpkm.TT.log.SD2k"",""polar"")])
+param = ""logFPKMsd2k_CNValldata_Polar""
+
+################################
+# SNF subtype 
+################################
+library(SNFtool)
+library(tidyr)
+library(CancerSubtypes)
+
+# First, set all the parameters:
+K = 15;		
+alpha = 0.5;  	
+T = 20; 	
+
+cnv.m.n = standardNormalization(cnv)
+exp.fpkm.m.n = standardNormalization(exp.fpkm.TT.log.SD2k)
+polar.m.n = standardNormalization(polar)
+
+# If the data is continuous, we recommend to use the function ""dist2"" as follows 
+dist_cnv = (SNFtool::dist2(cnv.m.n,cnv.m.n))
+dist_exp = (SNFtool::dist2(as.matrix(exp.fpkm.m.n),as.matrix(exp.fpkm.m.n)))
+dist_polar = (SNFtool::dist2(as.matrix(polar.m.n), as.matrix((polar.m.n))))
+
+# next, construct similarity graphs
+W_cnv = affinityMatrix(dist_cnv, K, alpha)
+W_exp = affinityMatrix(dist_exp, K, alpha)
+W_polar = affinityMatrix(dist_polar, K, alpha)
+
+W = SNF(list(W_cnv, W_exp, W_polar), K, T)
+#write.csv(W,file = paste0(param,""_W.csv"" ))
+
+
+################################
+# choose the best num of clusters
+################################
+library(Spectrum)
+library(plot3D)
+pdf(paste0(""./best_num_Spectral_cluster.pdf""))
+r <- Spectrum(W,method =1,diffusion = TRUE ,kerneltype = ""stsc"",NN = 20)
+dev.off()
+# optimal K: 4
+
+C = 4			# number of clusters
+set.seed(1234)
+group = spectralClustering(W, C) 	# the final subtypes information
+
+names(group) <- row.names(exp.fpkm.m.n)
+SNF_Cluster  <- group
+SNF_Cluster[group == 1] = 3
+SNF_Cluster[group == 2] = 1
+SNF_Cluster[group == 4] = 2
+SNF_Cluster[group == 3] = 4
+
+
+################################
+# estimate Silhouette width and visulize
+################################
+# displayClustersWithHeatmap
+set.seed(321)
+pdf(paste0(param,""_Re_nonRe_displayClustersWithHeatmap.C"",C,"".pdf""))
+displayClustersWithHeatmap(W = log(W), group = SNF_Cluster, 
+                           col = colorRampPalette(c(""#4F4F4F"",""#595959"",""#7F7F7F"",""#666666"",""#999999"",""#CCCCCC"",""#D9D9D9"",""#ECECEC""))(20),
+                           labRow = F,labCol = F) 
+dev.off()
+
+# Silhouette
+palette4silhou = c(""#3D76AE"",""#53AD4A"" ,""#EDAB3C"" ,""#C3392A"",rainbow(n = 5))
+pdf(paste0(param,""_Re_nonRe_silhouette_SimilarityMatrix.C"",C,"".pdf""))
+sil <- silhouette_SimilarityMatrix(SNF_Cluster, W)
+plot(sil, col = palette4silhou[1:C])
+dev.off()
+
+
+
+
+
+
+
+
+","R"