R语言ggboxplot-一文掌握箱线图绘制所有细节

2019-07-26 本文已影响49人医科研

如果没有时间精力学习代码，推荐了解：零代码数据挖掘课程

载入数据

Sys.setlocale('LC_ALL','C')
load(file = "F:/Bioinfor_project/Breast/AS_research/AS/result/hubgene.Rdata")
head(data)
require(cowplot)
require(tidyverse)
require(ggplot2)
require(ggsci)
require(ggpubr)
mydata<-data %>% 
  ## 基因表达数据gather,gather的范围应调整
  gather(key="gene",value="Expression",CCL14:TUBB3) %>% 
  ##
  dplyr::select(ID,gene,Expression,everything()) 
head(mydata)  ## 每个基因作为一个变量的宽数据

创建带有pvalue的箱线图

参考资料
展示绘图细节控制

p <- ggboxplot(mydata, x = "group", y = "Expression",
          color = "group", palette = "jama",
          add = "jitter")
#  Add p-value
p + stat_compare_means()

image.png

改变统计方法

# Change method
p + stat_compare_means(method = "t.test")

image.png

统计学意义标注

label="p.signif"
p.format等
label.x标注位置

p + stat_compare_means( label = "p.signif")

image.png

多组比较

给出global pvalue

# Default method = "kruskal.test" for multiple groups
ggboxplot(mydata, x = "gene", y = "Expression",
          color = "gene",add="jitter", palette = "jama")+
  stat_compare_means()

# Change method to anova
ggboxplot(mydata, x = "gene", y = "Expression",
          color = "gene", add="jitter", palette = "jama")+
  stat_compare_means(method = "anova")

image.png

指定比较

配对比较:会完成各个变量的比较，默认wilcox.test法，可修改
my_comparisions:可以指定自己想要进行的比较
指定参考组，进行比较

require(ggpubr)
compare_means(Expression ~ gene,  data = mydata)

## 指定自己想要的比较
# Visualize: Specify the comparisons you want
my_comparisons <- list( c("CCL14", "HBA1"), c("HBA1", "CCL16"), c("CCL16", "TUBB3") )
ggboxplot(mydata, x = "gene", y = "Expression",
          color = "group",add = "jitter", palette = "jama")+ 
  stat_compare_means(comparisons = my_comparisons)#+ # Add pairwise comparisons p-value
  #stat_compare_means()     # Add global p-value

image.png

指定参考组

指定CCL14作为参考组与其它各组比较
ref.group

compare_means(Expression ~ gene,  data = mydata, ref.group = "CCL14",
              method = "t.test")
# Visualize
mydata %>% 
  filter(group=="TNBC") %>% # 筛选TNBC数据
ggboxplot( x = "gene", y = "Expression",
          color = "gene",add = "jitter", palette = "nejm")+
  stat_compare_means(method = "anova")+      # Add global p-value
  stat_compare_means(label = "p.signif", method = "t.test",
                     ref.group = "CCL14")

image.png

多基因分面

按另外一个变量分组比较

## 比较各个基因在TNBC与Normal表达
compare_means( Expression ~ group, data = mydata, 
              group.by = "gene")
# Box plot facetted by "gene"
p <- ggboxplot(mydata, x = "group", y = "Expression",
          color = "group", palette = "jco",
          add = "jitter",
          facet.by = "gene", short.panel.labs = FALSE)
# Use only p.format as label. Remove method name.
p + stat_compare_means(label = "p.format")

image.png

将pvalue换成星号

hide.ns = TRUE.参数可隐藏ns

p + stat_compare_means(label =  "p.signif", label.x = 1.5)

image.png

将各个图绘制在一张图中

p <- ggboxplot(mydata, x = "gene", y = "Expression",
          color = "group", palette = "nejm",
          add = "jitter")
p + stat_compare_means(aes(group = group))

image.png

修改下pvalue展示的方式

# Show only p-value
p + stat_compare_means(aes(group = group), label = "p.format")

image.png

用星号表示pvalue

# Use significance symbol as label
p + stat_compare_means(aes(group = group), label = "p.signif")

image.png

配对样本比较

要求x,y具有相同的样本数，进行一一配对比较

head(ToothGrowth)
compare_means(len ~ supp, data = ToothGrowth, 
              group.by = "dose", paired = TRUE)
# Box plot facetted by "dose"
p <- ggpaired(ToothGrowth, x = "supp", y = "len",
          color = "supp", palette = "jama", 
          line.color = "gray", line.size = 0.4,
          facet.by = "dose", short.panel.labs = FALSE)
# Use only p.format as label. Remove method name.
p + stat_compare_means(label = "p.format", paired = TRUE)

image.png

封装为函数命名为group_box

功能：已经选定的基因绘制箱线图
- 参数1：group分组变量，可以是自己所有感兴趣的变量
参数2：mydata为整理好的清洁数据,gene为长数据(gather版本)

head(mydata)
group_box<-function(group=group,data=mydata){
        p <- ggboxplot(mydata, x = "gene", y = "Expression",
          color = group, 
          palette = "nejm",
          add = "jitter")
p + stat_compare_means(aes(group = group))
}

## 
group_box(group="PAM50",data = mydata)

封装为函数命名为group_box

功能：已经选定的基因绘制箱线图
- 参数1：group分组变量，可以是自己所有感兴趣的变量
参数2：mydata为整理好的清洁数据,gene为长数据(gather版本)

head(mydata)
group_box<-function(group=group,data=mydata){
        p <- ggboxplot(mydata, x = "gene", y = "Expression",
          color = group, 
          palette = "nejm",
          add = "jitter")
p + stat_compare_means(aes(group = group))
}

## 
group_box(group="PAM50",data = mydata)

image.png

封装函数gene_box

目的功能：对感兴趣的基因绘制和分组绘制boxplot
注意这时使用的应该是基因的宽数据，因为涉及到单个基因作为变量

head(data)
usedata<-data
## 封装函数
gene_box<-function(gene="CCL14",group="group",data=usedata){
p <- ggboxplot(data, x = group, y = gene,
          ylab = sprintf("Expression of %s",gene),
          xlab = group,
          color = group, 
          palette = "nejm",
          add = "jitter")
p + stat_compare_means(aes(group = group))
}

gene_box(gene="CCL14")

image.png

牛刀小试

gene_box(gene="CCL16",group="PAM50")

image.png

批量绘制

目的功能：绘制任意基因，任意分组，批量绘制一气呵成了
封装函数+lapply批量绘制无敌
在lapply中的函数参数设置，不在原函数中，而是直接放置在lapply中
do.call中参数1为函数，+c()包含原函数的参数设置，同样参数设置不在原函数中

require(gridExtra)
head(data)

## 需要批量绘制的基因名
name<-colnames(data)[3:6]
## 批量绘图
p<-lapply(name,gene_box,group = "T_stage")
## 组图
do.call(grid.arrange,c(p,ncol=2))

image.png

本期的内容就到这里，我是老朋友白介素2，下期再见。

R语言ggboxplot-一文掌握箱线图绘制所有细节

载入数据

创建带有pvalue的箱线图

改变统计方法

统计学意义标注

多组比较

指定比较

指定参考组

多基因分面

将pvalue换成星号

将各个图绘制在一张图中

修改下pvalue展示的方式

用星号表示pvalue

配对样本比较

封装为函数命名为group_box

封装为函数命名为group_box

封装函数gene_box

牛刀小试

批量绘制

猜你喜欢

热点阅读