本期内容为[跟着NC学作图]--生存分析（Survival analysis）

生存分析是人类癌症分析中常用到的分析之一。通过对表达水平分类，分成“High”组合“low”组，再进行生存分析即可。

最近，我打算做一系列的生物信息学相关的教程，这个系列是付费课程。主要包括转录组的组装（有参和有参）、lncRNA的鉴定及功能预测、可移动mRNA和lncRNA的鉴定、以及[R语言可视化-精美图形绘制系列]。
--
代码部分：

导入数据

vpc.clinical <- readRDS("./Data/VPC_clinical.Rdata")

> vpc.clinical[1:5,1:10]
  Sample Duplicates GC   Length     Fails Total_sequences_after_trim group
411 V33-33_S5_L007_R1_001_val_1   36.01205 50 147.3785  8.333333                   61644068   VPC
395    V28_S3_L003_R1_001_val_1   73.26805 50 147.9611 25.000000                   57841221   VPC
357    V18_S7_L003_R1_001_val_1   83.90558 50 148.0132 16.666667                   82224327   VPC
399    V29_S4_L002_R1_001_val_1   65.04211 59 147.2557 25.000000                   73173742   VPC
439    V46_S5_L002_R1_001_val_1   50.52673 53 147.1128 16.666667                   66938216   VPC
    experiment group.cp simpleName
411       Jess      VPC     V33-33
395       Jess      VPC        V28
357       Jess      VPC        V18
399       Jess      VPC        V29

分析处理、分类

vpc.clinical$GR_RPKM <- vpc.gr$GR_RPKM[match(vpc.clinical$Sample,vpc.gr$Sample)]
survival.data.sub <- subset(vpc.clinical,duplicated ==0)
survival.data.sub$OS.time <- as.numeric(survival.data.sub$`Published_Days to death or last followup`)
survival.data.sub$OS <- survival.data.sub$Published_Censored...14
survival.data.sub$OS[which(survival.data.sub$OS=="Yes")] <- 1
survival.data.sub$OS[which(survival.data.sub$OS=="No")] <- 2
survival.data.sub$OS <- as.numeric(survival.data.sub$OS)
survival.data.sub$RFS.time <- as.numeric(survival.data.sub$`Published_Days to progression or last followup`)
survival.data.sub$RFS <- survival.data.sub$Published_Censored...12
table(survival.data.sub$RFS )
survival.data.sub$RFS[which(survival.data.sub$RFS=="Yes")] <- 1
survival.data.sub$RFS[which(survival.data.sub$RFS=="No")] <- 2
survival.data.sub$RFS <- as.numeric(survival.data.sub$RFS)

survival.data.sub$GR_RPKM_log <- log2(survival.data.sub$GR_RPKM+1)
survival.data.sub$cluster <- "Low"
survival.data.sub$cluster[which(survival.data.sub$GR_RPKM_log>0)] <- "High"

提取数据

survival.data.sub$cluster <- factor(survival.data.sub$cluster,levels=c("High","Low"))
table(survival.data.sub$cluster)

> table(survival.data.sub$cluster)
High  Low 
  16   37

survfit()函数分析

OS.fit <- survfit(Surv(OS.time, OS) ~ cluster, 
                  data = survival.data.sub)

绘图

ggsurvplot(OS.fit,size = 1,# change line size
           palette = c("#2E9FDF", "#E7B800"), # custom color palette
           conf.int = FALSE, # Add confidence interval
           legend.title = "GR",
           pval = TRUE, # Add p-value
           risk.table = TRUE, # Add risk table
           risk.table.col ="strata", # Risk table color by groups
           ylab="Overall survival",
           xlab="Time (days)",
           legend.labs=c("High","Low"))

同样的操作


RFS.fit <- survfit(Surv(RFS.time, RFS) ~ cluster, 
                   data = survival.data.sub)
ggsurvplot(RFS.fit,size = 1,# change line size
           palette = c("#2E9FDF", "#E7B800"), # custom color palette
           conf.int = FALSE, # Add confidence interval
           legend.title = "GR",
           pval = TRUE, # Add p-value
           risk.table = TRUE, # Add risk table
           risk.table.col ="strata", # Risk table color by groups
           ylab="Progression-free survival",
           xlab="Time (days)",
           legend.labs=c("High","Low"))