Posted on

这篇文章我们将讲述

  1. 如何简单比较两组或多组的平均值
  2. 如何自动化为ggplot添加p值和显著性标记,包括箱线图、点图、条形图、线图等等

准备

安装和导入所需要的R包

需要R包ggpubr,版本>0.1.3,该包提供了基于ggplot2包的论文发表级绘图。

  • 从CRAN安装:
install.packages("ggpubr")

 

  • 或者从GitHub上下载最新的开发版本:
if(!require(devtools)) install.packages("devtools")

devtools::install_github("kassambara/ggpubr")

 

  • 载入ggpubr:
library(ggpubr)

#> Loading required package: ggplot2

 

ggpubr的官方文档在http://www.sthda.com/english/rpkgs/ggpubr ## 样例数据集

数据:ToothGrowth数据集

data("ToothGrowth")

head(ToothGrowth)

#> len supp dose

#> 1 4.2 VC 0.5

#> 2 11.5 VC 0.5

#> 3 7.3 VC 0.5

#> 4 5.8 VC 0.5

#> 5 6.4 VC 0.5

#> 6 10.0 VC 0.5

 

比较均值的方法

http://www.sthda.com/english/wiki/comparing-means-in-r包含了均值方法比较的详细描述,这里我们汇总常见的均值比较方法:

方法

R 函数

描述

T检验

t.test()

比较两组 (参数)

Wilcoxon检验

wilcox.test()

比较两组 (非参数)

ANOVA

aov() or anova()

比较多组 (参数)

Kruskal-Wallis

kruskal.test()

比较多组 (非参数)

添加p值的函数

这里我们展示ggpubr包中可以使用的用于添加p值的R函数:

  • compare_means()
  • stat_compare_means()

compare_mean()

下一部分我们将实际学习使用,这里先具体介绍一下这个函数:

compare_means(formula, data, method="wilcox.test", paired=FALSE,

group.by=NULL, ref.group = NULL, ...)

 

  • formula: x~group形式的公式,x是一个数值向量,group是有1个或者多个组别的因子。比如formula = TP53 ~ cancer_group。也可以使用多个响应变量,比如formula = c(TP53, PTEN) ~ cancer_group。
  • data: 包含变量的数据框
  • method: 检验的类型。默认是wilcox.test。允许的值包括:
    • t.test和wilcox.test。
    • anova和kruskal.test。
  • paired: 逻辑值,是否执行配对检验。仅能用于t.test和wilcox.test。
  • group.by: 在执行检验前对数据集进行分组的变量。指定后,会根据该变量分不同子集进行检验。
  • ref.group: 一个字符串,指定参考组。指定后,对于一个给定分组变量,每个分组水平都会和参考组进行比较。ref.group可以使用.all,这会对所有组别基于一个全局的均值进行两两比较。 ## stat_compare_means()

这个函数扩展了ggplot2,可以对指定ggplot图形添加均值比较的p值。

简单形式如下:

stat_compare_means(mapping = NULL,

comparisons = NULL,

hide.ns = FALSE,

label = NULL,

label.x = NULL,

label.y = NULL)

 

  • mapping: 由aes()创建的映射集合
  • comparisons: 一个长度为2的向量列表。向量中元素都是x轴的两个名字或者2个对于感兴趣,要进行比较的整数索引
  • hide.ns: 逻辑值,如果TRUE,隐藏不显著标记ns
  • label: 指定标签类型的字符串。允许值包括p.signif(显示显著性水平),p.format(显示p值)
  • label.x, label.y: 数值。用于摆放标签位置的坐标,如果太短,会循环重复。
  • …: 其他传入compare_means()的参数,例如method,paired,ref.group

比较两个独立组别

执行检验:

compare_means(len ~ supp, data = ToothGrowth)

#> # A tibble: 1 x 8

#> .y. group1 group2 p p.adj p.format p.signif method

#> <chr> <chr> <chr> <dbl> <dbl> <chr> <chr> <chr>

#> 1 len OJ VC 0.0645 0.064 0.064 ns Wilcoxon

 

默认执行method=”wilcox.test”,你可以指定method = “t.test”进行t检验。 返回一个包含下面列的数据框:

  • .y: 用于检验的y变量
  • p: p值
  • p.adj: 矫正p值,默认矫正方法p.adjust.method=”holm”
  • p.format: 格式化p值
  • p.signif: 显著性水平
  • 方法: 用于比较组别的统计检验

创建一个带p值的箱线图:

p <- ggboxplot(ToothGrowth, x="supp",

y = "len", color = "supp",

palette = "jco", add = "jitter")

# 添加p值

p + stat_compare_means()

 

https://shixiangwang.github.io/home/cn/post/2019-09-23-ggpubr-add-p-values-and-significance-levels_files/figure-html/unnamed-chunk-8-1.png

# 改变方法

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

 

https://shixiangwang.github.io/home/cn/post/2019-09-23-ggpubr-add-p-values-and-significance-levels_files/figure-html/unnamed-chunk-8-2.png

注意,p值标签位置可以使用label.x, label.y, hjust, vjust参数进行矫正。

默认p值标签显示是将compare_means()返回数据框的method和p列粘连到一起。你可以使用aes()函数进行组合。

比如,

  • aes(label = ..p.format..) 或者 aes(label = paste0(“p = “, ..p.format..)): 显示p值但不显示方法
  • aes(label = ..p.signif..): 仅显示显著性水平
  • aes(label = paste0(..method.., “\n”, “p =”, ..p.format..)) 使用在方法和p值中插入换行符

实际操作看看:

p + stat_compare_means(aes(label = ..p.signif..),

label.x = 1.5,

label.y = 40)

 

https://shixiangwang.github.io/home/cn/post/2019-09-23-ggpubr-add-p-values-and-significance-levels_files/figure-html/unnamed-chunk-9-1.png

也可以使用字符串指定标签:

p + stat_compare_means(label = "p.signif", label.x = 1.5, label.y = 40)

 

https://shixiangwang.github.io/home/cn/post/2019-09-23-ggpubr-add-p-values-and-significance-levels_files/figure-html/unnamed-chunk-10-1.png

比较两组配对样本

执行检验:

compare_means(len ~ supp, data = ToothGrowth,

paired = TRUE)

#> # A tibble: 1 x 8

#> .y. group1 group2 p p.adj p.format p.signif method

#> <chr> <chr> <chr> <dbl> <dbl> <chr> <chr> <chr>

#> 1 len OJ VC 0.00431 0.0043 0.0043 ** Wilcoxon

 

使用ggpaired()函数可视化:

ggpaired(ToothGrowth, x="supp", y="len",

color="supp", line.color="gray",

line.size=0.4, palette = "jco") +

stat_compare_means(paired = TRUE)

 

https://shixiangwang.github.io/home/cn/post/2019-09-23-ggpubr-add-p-values-and-significance-levels_files/figure-html/unnamed-chunk-12-1.png

比较多组样本

  • 全局检验
compare_means(len ~ dose, data = ToothGrowth,

method = "anova")

#> # A tibble: 1 x 6

#> .y. p p.adj p.format p.signif method

#> <chr> <dbl> <dbl> <chr> <chr> <chr>

#> 1 len 9.53e-16 9.50e-16 9.5e-16 **** Anova

 

用全局p值画图:

# 默认方法为 method = "kruskal.test"

ggboxplot(ToothGrowth, x = "dose", y = "len",

color = "dose", palette = "jco") +

stat_compare_means()

 

https://shixiangwang.github.io/home/cn/post/2019-09-23-ggpubr-add-p-values-and-significance-levels_files/figure-html/unnamed-chunk-14-1.png

# 将方法改为anova

ggboxplot(ToothGrowth, x = "dose", y = "len",

color = "dose", palette = "jco") +

stat_compare_means(method = "anova")

 

https://shixiangwang.github.io/home/cn/post/2019-09-23-ggpubr-add-p-values-and-significance-levels_files/figure-html/unnamed-chunk-14-2.png

  • 成对比较,如果分组变量包含超过两个水平,配对检验会自动执行。
# 执行成对比较

compare_means(len ~ dose, data = ToothGrowth)

#> # A tibble: 3 x 8

#> .y. group1 group2 p p.adj p.format p.signif method

#> <chr> <chr> <chr> <dbl> <dbl> <chr> <chr> <chr>

#> 1 len 0.5 1 0.00000702 0.000014 7.0e-06 **** Wilcoxon

#> 2 len 0.5 2 0.0000000841 0.00000025 8.4e-08 **** Wilcoxon

#> 3 len 1 2 0.000177 0.00018 0.00018 *** Wilcoxon

# 可视化: 指定你想要比较的组别

my_comparisons <- list(c("0.5","1"), c("1", "2"),

c("0.5", "2"))

ggboxplot(ToothGrowth, x="dose", y="len",

color="dose", palette = "jco") +

stat_compare_means(comparisons = my_comparisons) + #添加成对p值

stat_compare_means(label.y = 50) # 添加全局p值

#> Warning in wilcox.test.default(c(4.2, 11.5, 7.3, 5.8, 6.4, 10, 11.2, 11.2, :

#> cannot compute exact p-value with ties

#> Warning in wilcox.test.default(c(4.2, 11.5, 7.3, 5.8, 6.4, 10, 11.2, 11.2, :

#> cannot compute exact p-value with ties

#> Warning in wilcox.test.default(c(16.5, 16.5, 15.2, 17.3, 22.5, 17.3, 13.6, :

#> cannot compute exact p-value with ties

 

https://shixiangwang.github.io/home/cn/post/2019-09-23-ggpubr-add-p-values-and-significance-levels_files/figure-html/unnamed-chunk-16-1.png

如果你想要指定精确的y轴位置,使用label.y参数:

ggboxplot(ToothGrowth, x="dose", y="len",

color="dose", palette = "jco") +

stat_compare_means(comparisons = my_comparisons,

label.y=c(29, 35, 40)) +

stat_compare_means(label.y = 45)

#> Warning in wilcox.test.default(c(4.2, 11.5, 7.3, 5.8, 6.4, 10, 11.2, 11.2, :

#> cannot compute exact p-value with ties

#> Warning in wilcox.test.default(c(4.2, 11.5, 7.3, 5.8, 6.4, 10, 11.2, 11.2, :

#> cannot compute exact p-value with ties

#> Warning in wilcox.test.default(c(16.5, 16.5, 15.2, 17.3, 22.5, 17.3, 13.6, :

#> cannot compute exact p-value with ties

 

https://shixiangwang.github.io/home/cn/post/2019-09-23-ggpubr-add-p-values-and-significance-levels_files/figure-html/unnamed-chunk-17-1.png

  • 基于参考组的多重成对比较
# 基于参考组

compare_means(len ~ dose, data = ToothGrowth,

ref.group = "0.5",

method = "t.test")

#> # A tibble: 2 x 8

#> .y. group1 group2 p p.adj p.format p.signif method

#> <chr> <chr> <chr> <dbl> <dbl> <chr> <chr> <chr>

#> 1 len 0.5 1 1.27e- 7 1.30e- 7 1.3e-07 **** T-test

#> 2 len 0.5 2 4.40e-14 8.80e-14 4.4e-14 **** T-test

# 可视化

ggboxplot(ToothGrowth, x="dose", y="len",

color="dose", palette = "jco") +

stat_compare_means(method="anova", label.y=40) +

stat_compare_means(label="p.signif", method="t.test",

ref.group = "0.5")

 

https://shixiangwang.github.io/home/cn/post/2019-09-23-ggpubr-add-p-values-and-significance-levels_files/figure-html/unnamed-chunk-19-1.png

  • 以所有值为基准(base-mean)进行多个成对比较

如果出现很多组别,两两比较过于复杂,通过将所有数据汇总创建一个虚拟的样本,以它为基准进行比较。

# Comparison of each group against base-mean

compare_means(len ~ dose, data = ToothGrowth, ref.group = ".all.",

method = "t.test")

#> # A tibble: 3 x 8

#> .y. group1 group2 p p.adj p.format p.signif method

#> <chr> <chr> <chr> <dbl> <dbl> <chr> <chr> <chr>

#> 1 len .all. 0.5 0.000000290 0.00000087 2.9e-07 **** T-test

#> 2 len .all. 1 0.512 0.51 0.51 ns T-test

#> 3 len .all. 2 0.000000425 0.00000087 4.3e-07 **** T-test

 

可视化

ggboxplot(ToothGrowth, x = "dose", y = "len",

color = "dose", palette = "jco")+

stat_compare_means(method = "anova", label.y = 40)+ # Add global p-value

stat_compare_means(label = "p.signif", method = "t.test",

ref.group = ".all.") # Pairwise comparison against all

 

https://shixiangwang.github.io/home/cn/post/2019-09-23-ggpubr-add-p-values-and-significance-levels_files/figure-html/unnamed-chunk-21-1.png

这个方法有时候会非常有用,比如下面这个例子中,我们可以通过每个样本均值与所有样本的均值进行比较,判断基因水平是过表达还是下调了。

# Load myeloma data from GitHub

myeloma <- read.delim("https://raw.githubusercontent.com/kassambara/data/master/myeloma.txt")

# 执行检验

compare_means(DEPDC1 ~ molecular_group, data = myeloma,

ref.group = ".all.", method = "t.test")

#> # A tibble: 7 x 8

#> .y. group1 group2 p p.adj p.format p.signif method

#> <chr> <chr> <chr> <dbl> <dbl> <chr> <chr> <chr>

#> 1 DEPDC1 .all. Cyclin D-1 0.288 1.00e+0 0.29 ns T-test

#> 2 DEPDC1 .all. Cyclin D-2 0.424 1.00e+0 0.42 ns T-test

#> 3 DEPDC1 .all. MMSET 0.578 1.00e+0 0.58 ns T-test

#> 4 DEPDC1 .all. MAF 0.254 1.00e+0 0.25 ns T-test

#> 5 DEPDC1 .all. Hyperdiploid 0.0000000273 1.90e-7 2.7e-08 **** T-test

#> 6 DEPDC1 .all. Proliferation 0.0000239 1.20e-4 2.4e-05 **** T-test

#> 7 DEPDC1 .all. Low bone disease 0.00000526 3.20e-5 5.3e-06 **** T-test

 

可视化表达谱

ggboxplot(myeloma, x="molecular_group", y="DEPDC1",

color="molecular_group", add="jitter",

legend="none") +

rotate_x_text(angle = 45) +

geom_hline(yintercept = mean(myeloma$DEPDC1),

linetype=2) + # 添加base mean的水平线

stat_compare_means(method = "anova", label.y = 1600)+ # Add global annova p-value

stat_compare_means(label = "p.signif", method = "t.test",

ref.group = ".all.") # Pairwise comparison against all

 

https://shixiangwang.github.io/home/cn/post/2019-09-23-ggpubr-add-p-values-and-significance-levels_files/figure-html/unnamed-chunk-23-1.png

多个分组变量

  • 根据某个变量分组后两个独立样本的比较

执行检验:

compare_means(len ~ supp, data = ToothGrowth,

group.by = "dose")

#> # A tibble: 3 x 9

#> dose .y. group1 group2 p p.adj p.format p.signif method

#> <dbl> <chr> <chr> <chr> <dbl> <dbl> <chr> <chr> <chr>

#> 1 0.5 len OJ VC 0.0232 0.046 0.023 * Wilcoxon

#> 2 1 len OJ VC 0.00403 0.012 0.004 ** Wilcoxon

#> 3 2 len OJ VC 1 1 1.000 ns Wilcoxon

 

因为生成了不同的子图,根据变量分面

# 根据 “dose” 变量分面绘制箱线图

p <- ggboxplot(ToothGrowth, x = "supp", y = "len",

color = "supp", palette = "jco",

add = "jitter",

facet.by = "dose", short.panel.labs = FALSE)

# Use only p.format as label. Remove method name.

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

 

https://shixiangwang.github.io/home/cn/post/2019-09-23-ggpubr-add-p-values-and-significance-levels_files/figure-html/unnamed-chunk-25-1.png

# Or use significance symbol as label

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

 

https://shixiangwang.github.io/home/cn/post/2019-09-23-ggpubr-add-p-values-and-significance-levels_files/figure-html/unnamed-chunk-26-1.png

将这几个图绘制在单个面板内:

p <- ggboxplot(ToothGrowth, x = "dose", y = "len",

color = "supp", palette = "jco",

add = "jitter")

p + stat_compare_means(aes(group = supp))

 

https://shixiangwang.github.io/home/cn/post/2019-09-23-ggpubr-add-p-values-and-significance-levels_files/figure-html/unnamed-chunk-27-1.png

# 仅显示p值

p + stat_compare_means(aes(group = supp), label = "p.format")

 

https://shixiangwang.github.io/home/cn/post/2019-09-23-ggpubr-add-p-values-and-significance-levels_files/figure-html/unnamed-chunk-28-1.png

# 使用显著性标记

p + stat_compare_means(aes(group = supp), label = "p.signif")

 

https://shixiangwang.github.io/home/cn/post/2019-09-23-ggpubr-add-p-values-and-significance-levels_files/figure-html/unnamed-chunk-29-1.png

  • 分组后配对样本比较
compare_means(len ~ supp, data = ToothGrowth,

group.by = "dose", paired = TRUE)

#> # A tibble: 3 x 9

#> dose .y. group1 group2 p p.adj p.format p.signif method

#> <dbl> <chr> <chr> <chr> <dbl> <dbl> <chr> <chr> <chr>

#> 1 0.5 len OJ VC 0.0330 0.066 0.033 * Wilcoxon

#> 2 1 len OJ VC 0.0137 0.041 0.014 * Wilcoxon

#> 3 2 len OJ VC 1 1 1.000 ns Wilcoxon

 

可视化,按分组变量dose分面创建一个多面板箱线图:

p <- ggpaired(ToothGrowth, x="supp", y="len",

color="supp", palette = "jco",

line.color = "grey", 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)

 

https://shixiangwang.github.io/home/cn/post/2019-09-23-ggpubr-add-p-values-and-significance-levels_files/figure-html/unnamed-chunk-31-1.png

其他图形

  • 条形图和线图(一组变量)
# 条形图加均值标准误

ggbarplot(ToothGrowth, x = "dose", y = "len", add = "mean_se")+

stat_compare_means() + # Global p-value

stat_compare_means(ref.group = "0.5", label = "p.signif",

label.y = c(22, 29)) # compare to ref.group

 

https://shixiangwang.github.io/home/cn/post/2019-09-23-ggpubr-add-p-values-and-significance-levels_files/figure-html/unnamed-chunk-32-1.png

# 线图加均值标准误

ggline(ToothGrowth, x = "dose", y = "len", add = "mean_se")+

stat_compare_means() + # Global p-value

stat_compare_means(ref.group = "0.5", label = "p.signif",

label.y = c(22, 29))

 

https://shixiangwang.github.io/home/cn/post/2019-09-23-ggpubr-add-p-values-and-significance-levels_files/figure-html/unnamed-chunk-32-2.png

  • 条形图和线图(两组变量)
ggbarplot(ToothGrowth, x = "dose", y = "len", add = "mean_se",

color = "supp", palette = "jco",

position = position_dodge(0.8))+

stat_compare_means(aes(group = supp), label = "p.signif", label.y = 29)

 

https://shixiangwang.github.io/home/cn/post/2019-09-23-ggpubr-add-p-values-and-significance-levels_files/figure-html/unnamed-chunk-33-1.png

ggline(ToothGrowth, x = "dose", y = "len", add = "mean_se",

color = "supp", palette = "jco")+

stat_compare_means(aes(group = supp), label = "p.signif",

label.y = c(16, 25, 29))

 

https://shixiangwang.github.io/home/cn/post/2019-09-23-ggpubr-add-p-values-and-significance-levels_files/figure-html/unnamed-chunk-33-2.png

6 Replies to “基于ggpubr包为ggplot添加p值和显著性标记”

  1. 您好,我在使用stat_compare_means函数的时候有报错提示:Caused by error in `t.test.default()`:
    ! ‘y’观察值数量不够 这种情况该怎么解决呢 非常感谢!

    1. 尝试把stat_compare_means(comparison = list)中括号内的参数删掉,结果意外能算出来结果了……我先前在其他地方看到说comparison参数可以填入的向量内容可以为X轴的两个组别名(字符型),所以我先赋值list = list(“1″,”2”),为什么这样反而会有报错呢,求解

  2. 感谢分享,学习到了很多。
    但在使用wilcox.test计算显著性时显示wilcox 无法精确计算带连结的p值,按照网络上的教程解决,添加了exact=FALSE,但问题仍然没解决。请问是什么原因,可以拜托帮忙解答一下吗?
    代码:
    p <- plot_richness(qiimedata,"Group",
    measures = NULL,
    color="Group")+
    geom_boxplot(aes(fill=Group))+
    theme_bw()+xlab(NULL)+
    scale_color_aaas()+
    scale_fill_aaas(alpha = 0.7)
    library("ggpubr"); packageVersion("ggpubr")#事后检验分析
    mycompare=list(c("KO","WT"))
    p <- p + stat_compare_means(comparisons=mycompare,
    label = "p.signif",#显示显著性水平,p.format(显示p值)
    method = "wilcox.test",exact=FALSE)
    p
    报错:Warning: Computation failed in `stat_signif()`
    Caused by error in `if (scales$x$map(comp[1]) == data$group[1] | manual) …`:
    ! missing value where TRUE/FALSE needed
    Warning in wilcox.test.default(c(483, 172, 321, 203, 385, 237, 247, 522, :
    cannot compute exact p-value with ties

进哥哥进行回复 取消回复

邮箱地址不会被公开。 必填项已用*标注