Skip to content

GitLab

Commit cad0b7de authored by Marco Monti's avatar Marco Monti
Browse files

I updated the scripts

parent a78a5561
Hide whitespace changes
Inline Side-by-side
CB2025_figure_1_RNAseq.R
View file @ cad0b7de
library("readxl") library("readxl")
library("dplyr") library("dplyr")
#####Directories##### #####Directories#####
us <- "/Users/Squadrito/" # us <-"C:/Users/bresesti.chiara/"
us <-"C:/Users/bresesti.chiara/" # wdir1509<-paste0(us, "/Dropbox (HSR Global)/SquadritoM_1509_RNASeq_QIAseq_UPX/QIAseqUltraplexRNA_181342/primary_analysis")
wdir1509<-paste0(us,"/Dropbox (HSR Global)/SquadritoM_1509_RNASeq_QIAseq_UPX/QIAseqUltraplexRNA_181342/primary_analysis") # wdir1510<-paste0(us, "/Dropbox (HSR Global)/SquadritoM_1510_RNA_miRNA_QIAseq_UPX")
wdir1510<-paste0(us,"/Dropbox (HSR Global)/SquadritoM_1510_RNA_miRNA_QIAseq_UPX") # fdir <- paste0(us, "/Dropbox (HSR Global)/CancerGeneTherapy/Cancer Gene Therapy/MS/2024 Bresesti et al/Scripts/plots and tables used in figures")
fdir <- paste0(us,"/Dropbox (HSR Global)/CancerGeneTherapy/Cancer Gene Therapy/MS/2024 Bresesti et al/Scripts/plots and tables used in figures") #
# input_d <-paste0(us, "/Dropbox (HSR Global)/SquadritoM_1509_RNASeq_QIAseq_UPX/QIAseqUltraplexRNA_181342/primary_analysis")
# wdir1510 <-paste0(us, "/Dropbox (HSR Global)/SquadritoM_1510_RNA_miRNA_QIAseq_UPX")
#### Figure 1B #### # output_d <- paste0(us, "/Dropbox (HSR Global)/CancerGeneTherapy/Cancer Gene Therapy/MS/2024 Bresesti et al/Scripts/plots and tables used in figures")
setwd(wdir1509)
df1 <- as.data.frame(read_excel("QIAseqUltraplexRNA_181342.xlsx",sheet=3,col_names = T,skip = 1))#sheet: umis.genes.polyA-mouse input_dir <- "/beegfs/scratch/ric.squadrito/ric.squadrito/90-935462466_scRNAseq_Bresesti/Analysis_MM/GitLab_scripts/reference"
df1 <-df1[,-c(1,3:6)] #keep only gene name and UMI counts output_dir <- "/beegfs/scratch/ric.squadrito/ric.squadrito/90-935462466_scRNAseq_Bresesti/Analysis_MM/GitLab_scripts/Output"
df1[,-1] <-apply(df1[,-1],2,function(x){x/sum(x)*1000000}) #UMI normalized by CPM
gene_vector <- c("Cd19", "Ms4a1", "Fcer2a", "Ighm", "Cd8a", "Xcr1", "Itgae", "Itgax", ################################################################################
"Ccr2", "Itgam", "Mgl2", "Cd68", "Vcam1", "Csf1r", "Adgre1", # R script to generate data and figures for manuscript figures 1B, 1C, and 1D.
"Siglec1", "Hmox1", "Timd4", "Vsig4", "Clec4f", "Marco", "Pecam1", #
"Tek", "Lyve1", "Stab2") # This script reads RNA-seq and miRNA-seq data from Excel files, performs
df2 <- df1[df1$gene%in%gene_vector,] %>% arrange(factor(gene, levels=gene_vector)) # normalization and data transformation, and generates output tables and plots.
rownames(df2) <- df2[,1] #
df2.scaled <- as.data.frame(t(scale(t(df2[-1])))) #Zscore normalization. Scaled only works on columns, so need to transform # Figures generated/data produced:
setwd(fdir) # - Figure 1B: Heatmap data for selected marker genes (Excel table)
openxlsx::write.xlsx(df2.scaled,"Figure_1B_table.xlsx", rowNames=T) # - Figure 1C: Pairwise MA plots for miRNA data
#the df was exported and used to create a heatmap using Graphpad # - Figure 1D: Heatmap data for selected miRNA families (Excel table)
#
#### Figure 1C #### # Input data files:
# "edgeR_DGE_res_volcano.pdf" (in the 'wdir1509' folder) was imported in illustrator # - Figure 1B: "QIAseqUltraplexRNA_181342.xlsx" (Sheet 3: umis.genes.polyA-mouse)
# and merged with the plot generated by the code below # - Figure 1C & 1D: "173308.all_samples.summary.xlsx" (Sheet 2: miRNA_piRNA)
setwd(wdir1510) # - Figure 1D: "miRNA_Family.xlsx" (Sheet 1)
df1 <- as.data.frame(read_excel("173308.all_samples.summary.xlsx",sheet=2,col_names = T))#sheet: miRNA_piRNA #
df1 <-df1[,1:7]#UMI ################################################################################
rownames(df1)<-df1$miRNA
df2 <- apply(df1[,-1], 2, function(x) log(x)) #Log counts
df2 <- as.data.frame(limma::normalizeCyclicLoess(df2, weights = NULL, span=0.7, iterations = 5, method = "pairs")) #Cyclic loess normalization #### Figure 1B ####
colnames(df2)<-c("B cell","RPM","cDC1","cDC2","LSEC","KC") #Rename columns df1 <- as.data.frame(read_excel(paste0(input_dir, "/QIAseqUltraplexRNA_181342.xlsx"), sheet=3, col_names = T, skip=1)) #sheet: umis.genes.polyA-mouse
upper.panel<-function(x, y, ...){ df1 <-df1[,-c(1,3:6)] # keep only gene name and UMI counts
points(((x+y)/2),(x-y), cex=0.6, pch=19, col="grey") df1[,-1] <-apply(df1[,-1],2,function(x){x/sum(x)*1000000}) # UMI normalized by CPM
above <- (x-y-1)*((x+y)/2-5) > 5 & ((x+y)/2)>5 gene_vector <- c("Cd19", "Ms4a1", "Fcer2a", "Ighm", "Cd8a", "Xcr1", "Itgae", "Itgax",
points(((x+y)/2)[above],(x-y)[above], col="red", cex=0.6, pch=19) "Ccr2", "Itgam", "Mgl2", "Cd68", "Vcam1", "Csf1r", "Adgre1",
below <- (x-y+1)*((x+y)/2-5) < -5 & ((x+y)/2)>5 "Siglec1", "Hmox1", "Timd4", "Vsig4", "Clec4f", "Marco", "Pecam1",
points(((x+y)/2)[below],(x-y)[below], col="blue", cex=0.6, pch=19) "Tek", "Lyve1", "Stab2")
} #function for MA plot df2 <- df1[df1$gene%in%gene_vector,] %>% arrange(factor(gene, levels=gene_vector))
pairs(df2[,1:6], lower.panel = NULL, upper.panel = upper.panel, rownames(df2) <- df2[,1]
ylim=c(-8.5,8.5), xlim=c(2,14), cex.labels = 2) #pairwise plot df2.scaled <- as.data.frame(t(scale(t(df2[-1])))) # Zscore normalization. Scaled only works on columns, so need to transform
openxlsx::write.xlsx(df2.scaled, paste0(output_dir, "/Figure_1B_table.xlsx"), rowNames=T)
#### Figure 1D #### # the df was exported and used to create a heatmap using Graphpad
setwd(wdir1510)
df1 <- as.data.frame(read_excel("173308.all_samples.summary.xlsx",sheet=2,col_names = T))#sheet: miRNA_piRNA #### Figure 1C ####
df1 <- df1[-which(grepl("piR",df1[,1])),1:7]#UMI and piRNA removing # "edgeR_DGE_res_volcano.pdf" (in the 'wdir1509' folder) was imported in illustrator
df1[,1] <- gsub("/.*","",df1[,1]) #leave only first miRNA for ambiguous entries # and merged with the plot generated by the code below
df1[,-1] <- apply(df1[,-1],2,function(x){x/sum(x,na.rm=T)*1000000})#UMI normalized by CPM df1 <- as.data.frame(read_excel(paste0(input_dir, "/173308.all_samples.summary.xlsx"), sheet=2, col_names=T)) #sheet: miRNA_piRNA
df.families <- as.data.frame(read_excel("Analysis CB/miRNA Family.xlsx",sheet=1,col_names = T))#miRNA families from miRBase df1 <-df1[,1:7] #UMI
df.families <- df.families[df.families[,3]==10090,c(4,1)] #select mouse entries rownames(df1)<-df1$miRNA
df.families[,1] <- sub(pattern = "p.*",replacement ="p",x = df.families[,1]) df2 <- apply(df1[,-1], 2, function(x) log(x)) #Log counts
df2 <- merge(df.families,df1,by=1) df2 <- as.data.frame(limma::normalizeCyclicLoess(df2, weights = NULL, span=0.7, iterations = 5, method = "pairs")) #Cyclic loess normalization
df2 <- aggregate(df2[,-c(1:2)],by = df2["miR family"],FUN = sum,na.rm=T) #sum counts by family colnames(df2)<-c("B cell","RPM","cDC1","cDC2","LSEC","KC") #Rename columns
gene_vector <- c("miR-150-5p","miR-25-3p/32-5p/92-3p/363-3p/367-3p","miR-142-3p.1", upper.panel<-function(x, y, ...){
"miR-17-5p/20-5p/93-5p/106-5p","miR-191-5p", points(((x+y)/2),(x-y), cex=0.6, pch=19, col="grey")
"miR-15-5p/16-5p/195-5p/322-5p/497-5p","miR-26-5p","miR-138-5p", above <- (x-y-1)*((x+y)/2-5) > 5 & ((x+y)/2)>5
"miR-223-3p","miR-342-3p","miR-22-3p","miR-192-5p/215-5p", points(((x+y)/2)[above],(x-y)[above], col="red", cex=0.6, pch=19)
"miR-125-5p/351-5p","miR-126-3p.1","miR-199-3p") below <- (x-y+1)*((x+y)/2-5) < -5 & ((x+y)/2)>5
df3 <- subset(df2, df2$`miR family`%in% gene_vector) %>% arrange(factor(`miR family`, levels=gene_vector)) points(((x+y)/2)[below],(x-y)[below], col="blue", cex=0.6, pch=19)
rownames(df3) <- df3[,1] } # function for MA plot
df3.scaled <- as.data.frame(t(scale(t(df3[-1])))) #Zscore normalization. Scaled only works on columns, so need to transform pairs(df2[,1:6], lower.panel = NULL, upper.panel = upper.panel,
setwd(fdir) ylim=c(-8.5,8.5), xlim=c(2,14), cex.labels = 2) # pairwise plot
openxlsx::write.xlsx(df2.scaled,"Figure_1D_table.xlsx", rowNames=T)
#the df was exported and used to create a heatmap using Graphpad #### Figure 1D ####
df1 <- as.data.frame(read_excel(paste0(input_dir, "/173308.all_samples.summary.xlsx"), sheet=2, col_names=T)) #sheet: miRNA_piRNA
df1 <- df1[-which(grepl("piR",df1[,1])),1:7] # UMI and piRNA removing
df1[,1] <- gsub("/.*","",df1[,1]) # leave only first miRNA for ambiguous entries
df1[,-1] <- apply(df1[,-1],2,function(x){x/sum(x,na.rm=T)*1000000}) #UMI normalized by CPM
df.families <- as.data.frame(read_excel("Analysis CB/miRNA_Family.xlsx",sheet=1,col_names = T))#miRNA families from miRBase
df.families <- df.families[df.families[,3]==10090,c(4,1)] #select mouse entries
df.families[,1] <- sub(pattern = "p.*",replacement ="p",x = df.families[,1])
df2 <- merge(df.families,df1,by=1)
df2 <- aggregate(df2[,-c(1:2)],by = df2["miR family"],FUN = sum,na.rm=T) #sum counts by family
gene_vector <- c("miR-150-5p","miR-25-3p/32-5p/92-3p/363-3p/367-3p","miR-142-3p.1",
"miR-17-5p/20-5p/93-5p/106-5p","miR-191-5p",
"miR-15-5p/16-5p/195-5p/322-5p/497-5p","miR-26-5p","miR-138-5p",
"miR-223-3p","miR-342-3p","miR-22-3p","miR-192-5p/215-5p",
"miR-125-5p/351-5p","miR-126-3p.1","miR-199-3p")
df3 <- subset(df2, df2$`miR family`%in% gene_vector) %>% arrange(factor(`miR family`, levels=gene_vector))
rownames(df3) <- df3[,1]
df3.scaled <- as.data.frame(t(scale(t(df3[-1])))) #Zscore normalization. Scaled only works on columns, so need to transform
openxlsx::write.xlsx(df2.scaled, paste0(output_dir, "/Figure_1D_table.xlsx"), rowNames=T)
#the df was exported and used to create a heatmap using Graphpad
CB2025_figure_3_RNAseq.R
View file @ cad0b7de
library(readxl) library(readxl)
library(ggplot2) library(ggplot2)
library(ggrepel) library(ggrepel)
library(dplyr) library(dplyr)
library(fgsea) library(fgsea)
library(clusterProfiler) library(clusterProfiler)
library(enrichplot) library(enrichplot)
##### Directories ##### ##### Directories #####
#us <- "/Users/Squadrito/" # us <-"C:/Users/bresesti.chiara/"
us <-"C:/Users/bresesti.chiara/" # wdir<-paste0(us,"/Dropbox (HSR Global)/90-857433247_RNAseq_Squadrito/05-DGE-NoOut-Corr")
wdir<-paste0(us,"/Dropbox (HSR Global)/90-857433247_RNAseq_Squadrito/05-DGE-NoOut-Corr") # wdir_CB<-paste0(us, "/Dropbox (HSR Global)/90-857433247_RNAseq_Squadrito/Analysis CB_v2")
wdir_CB<-paste0(us, "/Dropbox (HSR Global)/90-857433247_RNAseq_Squadrito/Analysis CB_v2") # fdir <- paste0(us,"/Dropbox (HSR Global)/CancerGeneTherapy/Cancer Gene Therapy/MS/2024 Bresesti et al/Scripts/plots and tables used in figures")
fdir <- paste0(us,"/Dropbox (HSR Global)/CancerGeneTherapy/Cancer Gene Therapy/MS/2024 Bresesti et al/Scripts/plots and tables used in figures")
input_dir <- "/beegfs/scratch/ric.squadrito/ric.squadrito/90-935462466_scRNAseq_Bresesti/Analysis_MM/GitLab_scripts/reference"
#### Figure 3A&B #### output_dir <- "/beegfs/scratch/ric.squadrito/ric.squadrito/90-935462466_scRNAseq_Bresesti/Analysis_MM/GitLab_scripts/Output"
#Import df ################################################################################
setwd(wdir) #
miR_ctrl <- read_excel("edgeR_results.xlsx", sheet = "miR342-control") # R script to generate data and figures for manuscript figures 3A, 3B, 3C, and 3D.
sponge_ctrl <- read_excel("edgeR_results.xlsx", sheet = "spongeBT-control") # This script performs differential gene expression analysis and gene set enrichment
# analysis to generate volcano plots, empirical cumulative distribution function (ECDF) plots,
#Add DEG color and label for volcano plot # dot plots, and enrichment maps.
miR_ctrl$DEG <- "NO" #
miR_ctrl$DEG[miR_ctrl$logFC > 1 & miR_ctrl$PValue < 0.05] <- "UP" # Figures generated:
miR_ctrl$DEG[miR_ctrl$logFC < (-1) & miR_ctrl$PValue < 0.05] <- "DOWN" # - Figure 3A & 3B: Volcano plots of differentially expressed genes in miR-342-3p
miR_ctrl$DEG_label <- NA # overexpression and sponge experiments, and ECDF plots comparing
miR_ctrl$DEG_label[miR_ctrl$DEG != "NO"] <- miR_ctrl$...1[miR_ctrl$DEG != "NO"] # logFC distributions of all genes vs. miR-342-3p target genes.
# - Figure 3C: Dot plot visualizing enriched pathways from gene set enrichment analysis (GSEA).
sponge_ctrl$DEG <- "NO" # - Figure 3D: Enrichment map visualizing relationships between enriched pathways.
sponge_ctrl$DEG[sponge_ctrl$logFC > 1 & sponge_ctrl$PValue < 0.05] <- "UP" #
sponge_ctrl$DEG[sponge_ctrl$logFC < (-1) & sponge_ctrl$PValue < 0.05] <- "DOWN" # Input data files:
sponge_ctrl$DEG_label <- NA # - Figure 3A & 3B: "edgeR_results.xlsx" (Sheet "miR342-control" and "spongeBT-control")
sponge_ctrl$DEG_label[sponge_ctrl$DEG != "NO"] <- sponge_ctrl$...1[sponge_ctrl$DEG != "NO"] # (Output from differential gene expression analysis, likely using edgeR)
# - Figure 3A & 3B: "TargetScan8.0__miR-342-3p.predicted_targets.xlsx" (Sheet 1)
#Volcano plot (left panel) # (List of predicted miR-342-3p target genes from TargetScan)
data <- miR_ctrl # - Figure 3C & 3D: "EdgeR_results.xlsx" (Sheet "miR342-control")
ggplot(data=data, aes(x=logFC, y=-log10(PValue), col=DEG, label=DEG_label)) + # (Same as input for Figure 3A & 3B, used for GSEA)
geom_point() + # - Figure 3C & 3D: "miDB_sig5.MLS.rds" (RDS file containing gene sets for gene set enrichment analysis)
theme_minimal() + #
geom_text_repel() + ################################################################################
scale_color_manual(values=c("blue", "grey", "red")) +
geom_vline(xintercept=c(-1, 1), col="black") +
geom_hline(yintercept=-log10(0.05), col="black") +
scale_x_continuous(limits = c(-4.1,4.1), breaks = seq(-4,4,1)) + #### Figure 3A&B ####
scale_y_continuous(limits = c(0,20)) +
xlab(bquote(Log[2](FC))) + #Import df
ylab(bquote(-Log[10](PVal))) miR_ctrl <- read_excel(paste0(input_dir, "/edgeR_results.xlsx"), sheet = "miR342-control")
sponge_ctrl <- read_excel(paste0(input_dir, "/edgeR_results.xlsx"), sheet = "spongeBT-control")
data <- sponge_ctrl
ggplot(data=data, aes(x=logFC, y=-log10(PValue), col=DEG, label=DEG_label)) + #Add DEG color and label for volcano plot
geom_point() + miR_ctrl$DEG <- "NO"
theme_minimal() + miR_ctrl$DEG[miR_ctrl$logFC > 1 & miR_ctrl$PValue < 0.05] <- "UP"
geom_text_repel() + miR_ctrl$DEG[miR_ctrl$logFC < (-1) & miR_ctrl$PValue < 0.05] <- "DOWN"
scale_color_manual(values=c("blue", "grey", "red")) + miR_ctrl$DEG_label <- NA
geom_vline(xintercept=c(-1, 1), col="black") + miR_ctrl$DEG_label[miR_ctrl$DEG != "NO"] <- miR_ctrl$...1[miR_ctrl$DEG != "NO"]
geom_hline(yintercept=-log10(0.05), col="black") +
scale_x_continuous(limits = c(-4.1,4.1), breaks = seq(-4,4,1)) + sponge_ctrl$DEG <- "NO"
scale_y_continuous(limits = c(0,20)) + sponge_ctrl$DEG[sponge_ctrl$logFC > 1 & sponge_ctrl$PValue < 0.05] <- "UP"
xlab(bquote(Log[2](FC))) + sponge_ctrl$DEG[sponge_ctrl$logFC < (-1) & sponge_ctrl$PValue < 0.05] <- "DOWN"
ylab(bquote(-Log[10](PVal))) sponge_ctrl$DEG_label <- NA
sponge_ctrl$DEG_label[sponge_ctrl$DEG != "NO"] <- sponge_ctrl$...1[sponge_ctrl$DEG != "NO"]
#Import miR-342-3p target list (from TargetScan)
setwd(wdir_CB) #Volcano plot (left panel)
miR342_targets <- read_excel("TargetScan8.0__miR-342-3p.predicted_targets.xlsx") data <- miR_ctrl
miR342_targets <- filter(miR342_targets, miR342_targets$`Cumulative weighted context++ score`< (-0.3)) ggplot(data=data, aes(x=logFC, y=-log10(PValue), col=DEG, label=DEG_label)) +
geom_point() +
#ecdf plot (right panel) theme_minimal() +
data <- miR_ctrl geom_text_repel() +
test <-filter(data, data$...1 %in% miR342_targets$'Target gene') scale_color_manual(values=c("blue", "grey", "red")) +
plot(ecdf(data$logFC), lwd = 2, do.points=F, verticals=T, geom_vline(xintercept=c(-1, 1), col="black") +
ylab = "Fraction of DEG", geom_hline(yintercept=-log10(0.05), col="black") +
xlab = bquote(Log[2](FC)), scale_x_continuous(limits = c(-4.1,4.1), breaks = seq(-4,4,1)) +
xlim=c(-0.5,0.5), scale_y_continuous(limits = c(0,20)) +
ylim=c(0,1), xlab(bquote(Log[2](FC))) +
main="ECDF of gene expression: miR vs mut") + ylab(bquote(-Log[10](PVal)))
lines(ecdf(test$logFC), col= "blue", do.points=F, lwd = 2, verticals=T) +
abline(v=0, col="black") + data <- sponge_ctrl
abline(h=0.5, col="black") + ggplot(data=data, aes(x=logFC, y=-log10(PValue), col=DEG, label=DEG_label)) +
legend("bottomright", c("All genes","miR-342-3p targets"), geom_point() +
col = c("black","blue"), lwd=2, cex = 0.7) theme_minimal() +
ks.test(test$logFC, data$logFC, alternative = "g") #Kolmogorov-Smirnov test to calculate p-val of miR target distribution *not* greater than average data geom_text_repel() +
scale_color_manual(values=c("blue", "grey", "red")) +
data <- sponge_ctrl geom_vline(xintercept=c(-1, 1), col="black") +
test <-filter(data, data$...1 %in% miR342_targets$'Target gene') geom_hline(yintercept=-log10(0.05), col="black") +
plot(ecdf(data$logFC), lwd = 2, do.points=F, verticals=T, scale_x_continuous(limits = c(-4.1,4.1), breaks = seq(-4,4,1)) +
ylab = "Fraction of DEG", scale_y_continuous(limits = c(0,20)) +
xlab = bquote(Log[2](FC)), xlab(bquote(Log[2](FC))) +
xlim=c(-0.3,0.3), ylab(bquote(-Log[10](PVal)))
ylim=c(0,1),
main="ECDF of gene expression: sponge vs scr") + # Import miR-342-3p target list (from TargetScan)
lines(ecdf(test$logFC), col= "red", do.points=F, lwd = 2, verticals=T) + miR342_targets <- read_excel(paste0(input_dir, "/TargetScan8.0__miR-342-3p.predicted_targets.xlsx"))
abline(v=0, col="black") + miR342_targets <- filter(miR342_targets, miR342_targets$`Cumulative weighted context++ score`< (-0.3))
abline(h=0.5, col="black") +
legend("bottomright", c("All genes","miR-342-3p targets"), #ecdf plot (right panel)
col = c("black","red"), lwd=2, cex = 0.7) data <- miR_ctrl
ks.test(test$logFC, data$logFC, alternative = "l") #Kolmogorov-Smirnov test to calculate p-val of miR target distribution *not* less than average data test <-filter(data, data$...1 %in% miR342_targets$'Target gene')
plot(ecdf(data$logFC), lwd = 2, do.points=F, verticals=T,
#### Figure 3C&D #### ylab = "Fraction of DEG",
xlab = bquote(Log[2](FC)),
#Upload df xlim=c(-0.5,0.5),
setwd(wdir) ylim=c(0,1),
df1<- read_excel("EdgeR_results.xlsx",, sheet = "miR342-control") main="ECDF of gene expression: miR vs mut") +
setwd(wdir_CB) lines(ecdf(test$logFC), col= "blue", do.points=F, lwd = 2, verticals=T) +
miDB_sig5 <- readRDS("miDB_sig5.MLS.rds") #GO terms db abline(v=0, col="black") +
abline(h=0.5, col="black") +
#Rename pathways of interest in miDB_sig5 legend("bottomright", c("All genes","miR-342-3p targets"),
names(miDB_sig5)[names(miDB_sig5) == "HALLMARK_OXIDATIVE_PHOSPHORYLATION"] <- "Oxydative phosphorylation" col = c("black","blue"), lwd=2, cex = 0.7)
names(miDB_sig5)[names(miDB_sig5) == "GOBP_REGULATION_OF_CHOLESTEROL_METABOLIC_PROCESS"] <- "Regulation of cholesterol metabolic process" ks.test(test$logFC, data$logFC, alternative = "g") #Kolmogorov-Smirnov test to calculate p-val of miR target distribution *not* greater than average data
names(miDB_sig5)[names(miDB_sig5) == "GOBP_RESPONSE_TO_INTERLEUKIN_12"] <- "Response to IL12"
names(miDB_sig5)[names(miDB_sig5) == "GOBP_REGULATION_OF_LIPID_BIOSYNTHETIC_PROCESS"] <- "Regulation of lipid biosynthetic process" data <- sponge_ctrl
names(miDB_sig5)[names(miDB_sig5) == "GOMF_MHC_CLASS_I_PROTEIN_BINDING"] <- "MHC-I protein binding" test <-filter(data, data$...1 %in% miR342_targets$'Target gene')
names(miDB_sig5)[names(miDB_sig5) == "GOBP_TUMOR_NECROSIS_FACTOR_MEDIATED_SIGNALING_PATHWAY"] <- "TNFa mediated signalling pathway" plot(ecdf(data$logFC), lwd = 2, do.points=F, verticals=T,
names(miDB_sig5)[names(miDB_sig5) == "GOBP_NEGATIVE_REGULATION_OF_CELL_CYCLE_G2_M_PHASE_TRANSITION"] <- "Negative regulation of cell cycle progression" ylab = "Fraction of DEG",
names(miDB_sig5)[names(miDB_sig5) == "GOBP_FATTY_ACYL_COA_BIOSYNTHETIC_PROCESS"] <- "Fatty acyl-CoA biosynthetic process" xlab = bquote(Log[2](FC)),
names(miDB_sig5)[names(miDB_sig5) == "GOBP_RESPONSE_TO_TUMOR_NECROSIS_FACTOR"] <- "Response to TNFa" xlim=c(-0.3,0.3),
names(miDB_sig5)[names(miDB_sig5) == "HALLMARK_G2M_CHECKPOINT"] <- "G2M checkpoint" ylim=c(0,1),
names(miDB_sig5)[names(miDB_sig5) == "GOBP_RESPONSE_TO_TYPE_I_INTERFERON"] <- "Response to type I interferon" main="ECDF of gene expression: sponge vs scr") +
names(miDB_sig5)[names(miDB_sig5) == "GOBP_INTERLEUKIN_1_MEDIATED_SIGNALING_PATHWAY"] <- "IL1 mediated signalling pathway" lines(ecdf(test$logFC), col= "red", do.points=F, lwd = 2, verticals=T) +
names(miDB_sig5)[names(miDB_sig5) == "GOBP_RESPONSE_TO_INTERLEUKIN_1"] <- "Response to IL1" abline(v=0, col="black") +
names(miDB_sig5)[names(miDB_sig5) == "LPS_RO"] <- "LPS response genes" abline(h=0.5, col="black") +
names(miDB_sig5)[names(miDB_sig5) == "GOBP_CYTOKINE_MEDIATED_SIGNALING_PATHWAY"] <- "Cytokine mediated signalling pathway" legend("bottomright", c("All genes","miR-342-3p targets"),
names(miDB_sig5)[names(miDB_sig5) == "GOBP_MYELOID_CELL_DIFFERENTIATION"] <- "Myeloid cell differentiation" col = c("black","red"), lwd=2, cex = 0.7)
names(miDB_sig5)[names(miDB_sig5) == "GOBP_INTERLEUKIN_10_PRODUCTION"] <- "IL10 production" ks.test(test$logFC, data$logFC, alternative = "l") # Kolmogorov-Smirnov test to calculate p-val of miR target distribution *not* less than average data
names(miDB_sig5)[names(miDB_sig5) == "GOBP_POSITIVE_REGULATION_OF_ENDOTHELIAL_CELL_PROLIFERATION"] <- "Regulation of endothelial cell proliferation"
names(miDB_sig5)[names(miDB_sig5) == "GOBP_TUMOR_NECROSIS_FACTOR_SUPERFAMILY_CYTOKINE_PRODUCTION"] <- "TNFa superfamily cytokine production" #### Figure 3C&D ####
names(miDB_sig5)[names(miDB_sig5) == "HALLMARK_ANGIOGENESIS"] <- "Angiogenesis"
names(miDB_sig5)[names(miDB_sig5) == "GOMF_PATTERN_RECOGNITION_RECEPTOR_ACTIVITY"] <- "Pattern recognition receptor activity" # Load df
names(miDB_sig5)[names(miDB_sig5) == "PGE2_RO"] <- "PGE2 response genes" df1<- read_excel(paste0(input_dir, "/EdgeR_results.xlsx"), sheet = "miR342-control")
miDB_sig5 <- readRDS(paste0(input_dir, "/miDB_sig5.MLS.rds")) #GO terms db
#Filter low expression, NA and order by FC
df1 <- filter(df1, df1$logCPM>5 & !is.na(df1$...1)) #Rename pathways of interest in miDB_sig5
df1<-df1[order(df1$logFC),] names(miDB_sig5)[names(miDB_sig5) == "HALLMARK_OXIDATIVE_PHOSPHORYLATION"] <- "Oxydative phosphorylation"
df2<-df1$logFC names(miDB_sig5)[names(miDB_sig5) == "GOBP_REGULATION_OF_CHOLESTEROL_METABOLIC_PROCESS"] <- "Regulation of cholesterol metabolic process"
names(df2)<-df1$...1 names(miDB_sig5)[names(miDB_sig5) == "GOBP_RESPONSE_TO_INTERLEUKIN_12"] <- "Response to IL12"
names(miDB_sig5)[names(miDB_sig5) == "GOBP_REGULATION_OF_LIPID_BIOSYNTHETIC_PROCESS"] <- "Regulation of lipid biosynthetic process"
#Run GSEA with fgsea and filter by PVal names(miDB_sig5)[names(miDB_sig5) == "GOMF_MHC_CLASS_I_PROTEIN_BINDING"] <- "MHC-I protein binding"
sig <- miDB_sig5 names(miDB_sig5)[names(miDB_sig5) == "GOBP_TUMOR_NECROSIS_FACTOR_MEDIATED_SIGNALING_PATHWAY"] <- "TNFa mediated signalling pathway"
test1<-fgsea(sig, df2,minSize = 7,maxSize =500, nproc=1) names(miDB_sig5)[names(miDB_sig5) == "GOBP_NEGATIVE_REGULATION_OF_CELL_CYCLE_G2_M_PHASE_TRANSITION"] <- "Negative regulation of cell cycle progression"
PvalResult<-filter(test1, test1$padj<= 0.05) names(miDB_sig5)[names(miDB_sig5) == "GOBP_FATTY_ACYL_COA_BIOSYNTHETIC_PROCESS"] <- "Fatty acyl-CoA biosynthetic process"
names(miDB_sig5)[names(miDB_sig5) == "GOBP_RESPONSE_TO_TUMOR_NECROSIS_FACTOR"] <- "Response to TNFa"
#DotPlot pathways of interest (Fig.3C) names(miDB_sig5)[names(miDB_sig5) == "HALLMARK_G2M_CHECKPOINT"] <- "G2M checkpoint"
pathways <- c("Oxydative phosphorylation", names(miDB_sig5)[names(miDB_sig5) == "GOBP_RESPONSE_TO_TYPE_I_INTERFERON"] <- "Response to type I interferon"
"Regulation of cholesterol metabolic process", names(miDB_sig5)[names(miDB_sig5) == "GOBP_INTERLEUKIN_1_MEDIATED_SIGNALING_PATHWAY"] <- "IL1 mediated signalling pathway"
"Response to IL12", names(miDB_sig5)[names(miDB_sig5) == "GOBP_RESPONSE_TO_INTERLEUKIN_1"] <- "Response to IL1"
"Regulation of lipid biosynthetic process", names(miDB_sig5)[names(miDB_sig5) == "LPS_RO"] <- "LPS response genes"
"MHC-I protein binding", names(miDB_sig5)[names(miDB_sig5) == "GOBP_CYTOKINE_MEDIATED_SIGNALING_PATHWAY"] <- "Cytokine mediated signalling pathway"
"TNFa mediated signalling pathway", names(miDB_sig5)[names(miDB_sig5) == "GOBP_MYELOID_CELL_DIFFERENTIATION"] <- "Myeloid cell differentiation"
"Negative regulation of cell cycle progression", names(miDB_sig5)[names(miDB_sig5) == "GOBP_INTERLEUKIN_10_PRODUCTION"] <- "IL10 production"
"Fatty acyl-CoA biosynthetic process", names(miDB_sig5)[names(miDB_sig5) == "GOBP_POSITIVE_REGULATION_OF_ENDOTHELIAL_CELL_PROLIFERATION"] <- "Regulation of endothelial cell proliferation"
"Response to TNFa", names(miDB_sig5)[names(miDB_sig5) == "GOBP_TUMOR_NECROSIS_FACTOR_SUPERFAMILY_CYTOKINE_PRODUCTION"] <- "TNFa superfamily cytokine production"
"G2M checkpoint", names(miDB_sig5)[names(miDB_sig5) == "HALLMARK_ANGIOGENESIS"] <- "Angiogenesis"
"Response to type I interferon", names(miDB_sig5)[names(miDB_sig5) == "GOMF_PATTERN_RECOGNITION_RECEPTOR_ACTIVITY"] <- "Pattern recognition receptor activity"
"IL1 mediated signalling pathway", names(miDB_sig5)[names(miDB_sig5) == "PGE2_RO"] <- "PGE2 response genes"
"Response to IL1",
"LPS response genes", # Filter low expression, NA and order by FC
"Cytokine mediated signalling pathway", df1 <- filter(df1, df1$logCPM>5 & !is.na(df1$...1))
"Myeloid cell differentiation", df1<-df1[order(df1$logFC),]
"IL10 production", df2<-df1$logFC
"Regulation of endothelial cell proliferation", names(df2)<-df1$...1
"TNFa superfamily cytokine production",
"Angiogenesis", # Run GSEA with fgsea and filter by PVal
"Pattern recognition receptor activity", sig <- miDB_sig5
"PGE2 response genes") test1 <- fgsea(sig, df2, minSize=7, maxSize=500, nproc=1)
PvalResult <- filter(test1, test1$padj <= 0.05)
genelist <- as.data.frame(PvalResult[PvalResult$pathway %in% pathways,])
ggplot(genelist, aes(x=NES, y=reorder(pathway,NES), size=size ,color=padj)) + # DotPlot pathways of interest (Fig.3C)
geom_point() + pathways <- c("Oxydative phosphorylation",
scale_size_area(limits=c(10,450), max_size = 15) + "Regulation of cholesterol metabolic process",
scale_colour_gradient(low="red",high="blue") + "Response to IL12",
labs(y='Pathway',x='NES') "Regulation of lipid biosynthetic process",
"MHC-I protein binding",
#Run GSEA with ClusterProfiler "TNFa mediated signalling pathway",
genelist <- data.frame(term = rep(names(miDB_sig5), sapply(miDB_sig5, length)), "Negative regulation of cell cycle progression",
gene = unlist(miDB_sig5)) "Fatty acyl-CoA biosynthetic process",
df2 = sort(df2, decreasing = TRUE) "Response to TNFa",
test2 <- GSEA(df2,TERM2GENE = genelist) "G2M checkpoint",
"Response to type I interferon",
#Enrichment map (Fig.3D) "IL1 mediated signalling pathway",
test3 <- filter(test2, ID %in% pathways) "Response to IL1",
test3 <- pairwise_termsim(test3) "LPS response genes",
emapplot(test3, min_edge = 0.01, color = "NES", layout="fr", repel =T)+ "Cytokine mediated signalling pathway",
scale_fill_gradient2(name=bquote(NES),low="blue", high="red") "Myeloid cell differentiation",
#N.B. Produces slightly different plot every time, "IL10 production",
#but connection between pathways stays the same "Regulation of endothelial cell proliferation",
\ No newline at end of file "TNFa superfamily cytokine production",
"Angiogenesis",
"Pattern recognition receptor activity",
"PGE2 response genes")
genelist <- as.data.frame(PvalResult[PvalResult$pathway %in% pathways,])
ggplot(genelist, aes(x=NES, y=reorder(pathway,NES), size=size ,color=padj)) +
geom_point() +
scale_size_area(limits=c(10,450), max_size = 15) +
scale_colour_gradient(low="red",high="blue") +
labs(y='Pathway', x='NES')
#Run GSEA with ClusterProfiler
genelist <- data.frame(term = rep(names(miDB_sig5), sapply(miDB_sig5, length)),
gene = unlist(miDB_sig5))
df2 = sort(df2, decreasing = TRUE)
test2 <- GSEA(df2,TERM2GENE = genelist)
#Enrichment map (Fig.3D)
test3 <- filter(test2, ID %in% pathways)
test3 <- pairwise_termsim(test3)
emapplot(test3, min_edge = 0.01, color = "NES", layout="fr", repel =T) +
scale_fill_gradient2(name=bquote(NES),low="blue", high="red")
\ No newline at end of file
CB2025_figure_5_scRNAseq.R
View file @ cad0b7de
...@@ -29,7 +29,28 @@ dir.create(plot_dir, showWarnings=F, recursive=T) ...@@ -29,7 +29,28 @@ dir.create(plot_dir, showWarnings=F, recursive=T)
sig <- readRDS("/beegfs/scratch/ric.squadrito/ric.squadrito/90-935462466_scRNAseq_Bresesti/reference/miDB_sig5.MLS.rds") sig <- readRDS("/beegfs/scratch/ric.squadrito/ric.squadrito/90-935462466_scRNAseq_Bresesti/reference/miDB_sig5.MLS.rds")
############################################################################### ################################################################################
# R script for single-cell RNA-seq data analysis and figure generation.
#
# This script performs cell annotation, visualization, differential gene expression
# analysis, and gene set enrichment analysis (GSEA) on single-cell RNA-seq data.
#
# Figures generated/data produced:
# - Figure 5A: UMAP visualization of cell clusters with final annotation
# - Figure 5B: UMAP plot with overlayed density of mOrange+ cells
# - Figure 5C: Dotplot of Slc7a11 gene expression across cell types and groups
# - Figure 5D: Barplot of GSEA results for selected gene sets
# - Figure 5E: Heatmap of cytokine signature GSEA results
# - Supplementary Figure 5A: Dotplot of marker gene expression across cell types
# - Supplementary Figure 5B: CSV tables for cell distribution per cluster and sample
# - Supplementary Figure 5C: CSV tables for mOrange+ cell distribution per cluster and sample
#
# Input data files:
# - RDS object: "CB1_CB3_CB4_final.rds" (Seurat object containing scRNA-seq data)
# - RDS object: "miDB_sig5.MLS.rds" (GO terms database for GSEA)
#
################################################################################
set.seed(42) set.seed(42)
......
TCGA_analysis.R
View file @ cad0b7de
...@@ -7,44 +7,73 @@ library(readxl) ...@@ -7,44 +7,73 @@ library(readxl)
library(survminer) library(survminer)
library(gridExtra) library(gridExtra)
library(ggplot2) library(ggplot2)
##Load dataset ##Load dataset
#username <- "C:/Users/notaro.marco/" wdir <- "/beegfs/scratch/ric.squadrito/ric.squadrito/90-935462466_scRNAseq_Bresesti/Analysis_MM/GitLab_scripts"
username <- "/Users/bresesti.chiara/"
wdir<- paste0(username, "/Dropbox (HSR Global)/CancerGeneTherapy/Cancer Gene Therapy/MS/2024 Bresesti et al Cell Reports/TCGA_analysis")
setwd(wdir) setwd(wdir)
metaD<-data.frame(fread("TCGA_phenotype_denseDataOnlyDownload.tsv.gz", sep='\t',colClasses=c("character"),data.table=FALSE))
Surv01<-data.frame(fread("Survival_SupplementalTable_S1_20171025_xena_sp", sep='\t',colClasses=c("character"),data.table=FALSE)) input_dir <- "/beegfs/scratch/ric.squadrito/ric.squadrito/90-935462466_scRNAseq_Bresesti/Analysis_MM/GitLab_scripts/reference"
output_dir <- "/beegfs/scratch/ric.squadrito/ric.squadrito/90-935462466_scRNAseq_Bresesti/Analysis_MM/GitLab_scripts/Output"
################################################################################
# R script for TCGA survival analysis of miR-342-3p expression.
#
# This script analyzes TCGA pancancer miRNA expression and survival data to assess
# the prognostic value of hsa-miR-342-3p across different tumor types.
# It calculates hazard ratios (HR) and generates Kaplan-Meier survival curves
# to visualize the association between miR-342-3p expression levels and patient survival.
#
# Figures generated:
# - HR_alltumors_342.pdf (HR forest plot): Forest plot visualizing hazard ratios and
# significance of miR-342-3p expression on overall survival across various tumor types.
# - Survival_***_342 (Survival curves - multiple tumors): Set of Kaplan-Meier survival plots
# for tumor types showing significant hazard ratios, illustrating survival differences
# between patients with high and low miR-342-3p expression.
# - Survival_metastatic_342.pdf (Survival curve - metastatic tumors): Kaplan-Meier survival plot
# specifically for metastatic tumors, showing the impact of miR-342-3p expression on survival
# in this patient subgroup.
#
# Input data files:
# - TCGA_phenotype_denseDataOnlyDownload.tsv.gz: TCGA patient phenotype data (metadata),
# downloaded from TCGA or Xena, contains clinical and demographic information.
# - Survival_SupplementalTable_S1_20171025_xena_sp: TCGA patient survival data, provides overall survival (OS) time and status.
# - pancanMiRs_EBadjOnProtocolPlatformWithoutRepsWithUnCorrectMiRs_08_04_16.xena.gz:
# TCGA pancancer miRNA expression data (FPKM values), downloaded from Xena,
# contains miRNA expression levels across different tumor samples.
#
# Note:
# - The script filters tumor types based on p-value significance (p < 0.1 for HR plot, p < 0.05 for survival curves)
# and minimum patient number (n > 100 for HR plot). These thresholds can be adjusted within the script.
################################################################################
metaD<-data.frame(fread(paste0(input_dir, "/TCGA_phenotype_denseDataOnlyDownload.tsv.gz"), sep='\t', colClasses=c("character"), data.table=FALSE))
Surv01<-data.frame(fread(paste0(input_dir, "/Survival_SupplementalTable_S1_20171025_xena_sp"), sep='\t', colClasses=c("character"), data.table=FALSE))
# FPKM01<-fread("tcga_RSEM_Hugo_norm_count") # FPKM01<-fread("tcga_RSEM_Hugo_norm_count")
FPKM01<-fread("pancanMiRs_EBadjOnProtocolPlatformWithoutRepsWithUnCorrectMiRs_08_04_16.xena.gz") FPKM01<-fread(paste0(input_dir, "/pancanMiRs_EBadjOnProtocolPlatformWithoutRepsWithUnCorrectMiRs_08_04_16.xena.gz"))
FPKM01<-as.data.frame(FPKM01) FPKM01<-as.data.frame(FPKM01)
# Annot<-data.frame(fread("probeMap_gencode.v23.annotation.gene.probemap", sep='\t',colClasses=c("character"),data.table=FALSE)) # Annot<-data.frame(fread(paste0(input_dir, "/probeMap_gencode.v23.annotation.gene.probemap"), sep='\t',colClasses=c("character"),data.table=FALSE))
allGenes <- FPKM01[,1] allGenes <- FPKM01[,1]
####################################################################### #######################################################################
#####PLOT HR Score of signature for multiple tumor types############### #####PLOT HR Score of signature for multiple tumor types###############
##################################################################### #####################################################################
humanGenes<-list(c("hsa-miR-342-3p")) humanGenes<-list(c("hsa-miR-342-3p"))
####create xlsx file with genes, patients and tumor types ####create xlsx file with genes, patients and tumor types
Pan.data3 <- metaD[,] Pan.data3 <- metaD[,]
Pan.data4 <- merge(Surv01,Pan.data3,by=1) Pan.data4 <- merge(Surv01,Pan.data3,by=1)
Genes_selected <- FPKM01[FPKM01[,1]%in%humanGenes,] Genes_selected <- FPKM01[FPKM01[,1]%in%humanGenes,]
rownames(Genes_selected)<-Genes_selected[,1] rownames(Genes_selected)<-Genes_selected[,1]
Genes_selected<-Genes_selected[,-1] Genes_selected<-Genes_selected[,-1]
Genes_selected <- data.frame(t(Genes_selected)) Genes_selected <- data.frame(t(Genes_selected))
Genes_selected$sample<-rownames(Genes_selected) Genes_selected$sample<-rownames(Genes_selected)
Pan.data5 <- merge(Genes_selected,Pan.data4,by.y="sample") Pan.data5 <- merge(Genes_selected,Pan.data4,by.y="sample")
colnames(Pan.data5) colnames(Pan.data5)
#setwd(wdir) write.xlsx(Pan.data5, paste0(ourdir, "/miR342_patients_survival.xlsx"))
#write.xlsx(Pan.data5, "miR342_patients_survival.xlsx")
#############Regression with defined k2 #############Regression with defined k2
...@@ -63,7 +92,7 @@ for (i in unique(Pan.data5$cancer.type.abbreviation)){ ...@@ -63,7 +92,7 @@ for (i in unique(Pan.data5$cancer.type.abbreviation)){
df2 <- rbind(df2,df1) df2 <- rbind(df2,df1)
df2<-df2[(order(df2$p)),] df2<-df2[(order(df2$p)),]
}else{}} }else{}}
#write.xlsx(df2, "miR342_HR_by_tumortype.xlsx") write.xlsx(df2, paste0(output_dir, "/miR342_HR_by_tumortype.xlsx"))
###Filters selected tumors ###Filters selected tumors
...@@ -75,21 +104,13 @@ a<-ggplot(df3, aes(x = reorder(Tumor, HR), y = HR, color = significance, label = ...@@ -75,21 +104,13 @@ a<-ggplot(df3, aes(x = reorder(Tumor, HR), y = HR, color = significance, label =
geom_point(size = 4) + geom_point(size = 4) +
geom_errorbar(aes(ymin = pmax(HR-SE, 0), ymax = HR + SE), width = 0.2) + geom_errorbar(aes(ymin = pmax(HR-SE, 0), ymax = HR + SE), width = 0.2) +
scale_color_manual(name = "Color", values = c("red", "blue")) + scale_color_manual(name = "Color", values = c("red", "blue")) +
labs(title = "HR by tumor type", labs(title = "HR by tumor type", x = "Tumor Type", y = "HR") +
x = "Tumor Type",
y = "HR") +
geom_rect(aes(xmin = -Inf, xmax = Inf, ymin = -Inf, ymax = 1), fill = "lightblue", alpha = 0.01) + geom_rect(aes(xmin = -Inf, xmax = Inf, ymin = -Inf, ymax = 1), fill = "lightblue", alpha = 0.01) +
geom_rect(aes(xmin = -Inf, xmax = Inf, ymin = 1, ymax = Inf), fill = "pink", alpha = 0.01)+ geom_rect(aes(xmin = -Inf, xmax = Inf, ymin = 1, ymax = Inf), fill = "pink", alpha = 0.01)+
theme_minimal() + theme_minimal() +
theme(axis.text.x = element_text(angle = 45, hjust= 1, color = "Black")) + theme(axis.text.x = element_text(angle = 45, hjust= 1, color = "Black")) +
geom_hline(yintercept = 1, linetype = "dashed", color = "gray") geom_hline(yintercept = 1, linetype = "dashed", color = "gray")
ggsave(filename = paste0(output_dir, "/HR_alltumors_342.pdf"), plot=a, width=9, height=4)
# pdf("HR_alltumors_342.pdf",width=9,height = 4)
a
# dev.off()
####Plot survival of chosen tumors ####Plot survival of chosen tumors
...@@ -98,7 +119,7 @@ df3<-df3[order(df3$HR),] ...@@ -98,7 +119,7 @@ df3<-df3[order(df3$HR),]
df3<-df3[df3$p<0.05,] df3<-df3[df3$p<0.05,]
# pdf("Survival_***_342",width=10,height = 6) #pdf(paste0(output_dir, "/Survival_***_342"), width=10,height = 6)
par(mfrow = c(2,4)) par(mfrow = c(2,4))
survival<-list() survival<-list()
for (i in df3$Tumor){ for (i in df3$Tumor){
...@@ -141,7 +162,7 @@ formatted_p_value <- ifelse(p_value < 0.0001, "<0.0001", as.numeric(round(p_val ...@@ -141,7 +162,7 @@ formatted_p_value <- ifelse(p_value < 0.0001, "<0.0001", as.numeric(round(p_val
par(las = 0) par(las = 0)
pdf("Survival_metastatic_342.pdf",width=5,height =5) pdf(paste0(output_dir, "/Survival_metastatic_342.pdf"), width=5,height =5)
plot(fit.score, lty = c(1, 1), col = c("blue", "red"), xlab = "Time (d)", ylab = "Overall Survival", lwd = 2, bty = "n", plot(fit.score, lty = c(1, 1), col = c("blue", "red"), xlab = "Time (d)", ylab = "Overall Survival", lwd = 2, bty = "n",
main = "Metastatic tumors") main = "Metastatic tumors")
#Add legend #Add legend
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment