上手即用，分组统计检验直方图的绘图脚本分享

# 读入表达矩阵
readFlie=function(input,type,row=T,header=T){
  # input 为读入文件的路径，type为读入文件的类型，格式为‘.txt’或‘.csv’,row=T,将文件的第一列设置为列名
  library(data.table,quietly = TRUE)
  if(type=='txt'){
    dat = fread(input,header = header,sep='\t',stringsAsFactors = F,check.names = F)
    if(row){
      dat = as.data.frame(dat,stringsAsFactors = F)
      rownames(dat) = dat[,1]
      dat = dat[,-1]
    }else{
      dat = as.data.frame(dat,stringsAsFactors = F)
    }
  }else{
    dat = fread(input,header = header,sep=',',stringsAsFactors = F,check.names = F)
    if(row){
      dat = as.data.frame(dat,stringsAsFactors = F)
      rownames(dat) = dat[,1]
      dat = dat[,-1]
    }else{
      dat = as.data.frame(dat,stringsAsFactors = F)
    }
  }
  return(dat)
}

# 保存图片,只支持ggplot对象
savePlots=function(path,plot,type=c('pdf','png','tiff','svg')[1],width=10,height=8,dpi=300){
  # path表示保存图片路径,需要加上相应的文件扩展名称，plot表示ggplot对象，type表示要保存的图片格式
  if(type=='pdf'){
    ggplot2::ggsave(filename = path,plot = plot,width = width,height = height,device = 'pdf')
  }else if(type=='svg'){
    svglite::ggsave(filename = path,plot = plot,width = width,height = height,device = 'svg')
  }else if(type=='png'){
    ggplot2::ggsave(filename = path,plot = plot,width = width,height = height,device = 'png',dpi = dpi)
  }else{
    ggplot2::ggsave(filename = path,plot = plot,width = width,height = height,device = 'tiff',dpi = dpi)
  }
}
# 分组条形图
PlotGroup_barplot=function(data,xlab = '', ylab = 'Value',legend = 'Group',
                             color =NULL,show_compare= T,cmp_show_compare = T,
                             test_method = c('wilcox.test','t.test','kruskal.test','anova')[1],
                             cmp_test_method = c('wilcox.test','t.test')[1],
                             orientation = c('vertical','horizontal','reverse')[1],
                             p_show = c('p.signif','p.format')[1],
                             add = c( 'mean_se', 'mean_sd', 'mean_ci', 'mean_range', 'median','mean', 'median_iqr')[1]){
# data是一个数据框，包含三列，第一列是样本名称(X轴坐标)，第二列是样本分组信息，第三列为值(Y轴坐标)
  # xlab，ylab和labs 分别自定义X轴名称，Y轴名称和图例标题名称
  # colors 自定义颜色，默认为NULL，函数会根据RColorBrewer包进行配色
  # show_compare 是否展示同组样本之间所有组别之间差异比较
  # test_method 同组样本之间所有组别之间差异比较方法
  # cmp_show_compare T or F 是否展示同组样本之间两两进行比较,样本分组必须大于2
  # cmp_test_method 同组样本之间两两进行比较方法
  # p_show 图片上差异显著性展示方式，*: p <= 0.05 **: p <= 0.01 ***: p <= 0.001 ****: p <= 0.0001
  # add 在条形图上添加统计值如样本的标准差，标准误，详细参数见https://www.rdocumentation.org/packages/ggpubr/versions/0.4.0/topics/add_summary 
  library(ggplot2)
  library(ggpubr)
  library(rstatix)
  library(dplyr)
  if(is.null(colors)){
    library(RColorBrewer)
    color = c(brewer.pal(9,'Set1'),brewer.pal(8,'Set2'),brewer.pal(12,'Set3'),
              brewer.pal(9,'Pastel1'),brewer.pal(8,'Pastel2'),brewer.pal(8,'Accent'),
              brewer.pal(8,'Dark2'),brewer.pal(12,'Paired')) %>% unique()
  }
  names(data) = c('Sample','Group',"Value")
  # 绘图
  p = ggbarplot(data, x = "Sample", y = "Value",fill = "Group", xlab = xlab,ylab = ylab,add = add,width = 0.5,legend.title = legend,palette = color,position = position_dodge(0.6),ggtheme = theme_bw()) + scale_y_continuous(expand = expansion(mult = c(0, 0.1)))
  if(show_compare){
    p = p + stat_compare_means(aes(group = Group), label = p_show,method = test_method )
  }
  # 组内样本两两之间差异比较
  if(cmp_show_compare & length(unique(data$Group))>2){
    if(cmp_test_method=='wilcox.test'){
      stat.test = data %>%
        group_by(Sample) %>%
        wilcox_test(Value ~ Group,p.adjust.method = "none") 
      stat.test = stat.test %>%
        add_xy_position(x = "Sample", fun = "mean_sd", dodge = 0.8) %>% dplyr::filter(p<0.05)
    }else{
      stat.test = data %>%
        group_by(Sample) %>%
        t_test(Value ~ Group,p.adjust.method = "none") 
      stat.test = stat.test %>%
        add_xy_position(x = "Sample", fun = "mean_sd", dodge = 0.8) %>% dplyr::filter(p<0.05)
    }
    if(p_show == 'p.signif'){
      p = p + stat_pvalue_manual(stat.test, label = "p.adj.signif", tip.length = 0.005,bracket.nudge.y = -2)
    }else{
      p = p + stat_pvalue_manual(stat.test, label = "p.adj", tip.length = 0.005,bracket.nudge.y = -2)
    }
  }
  return(p)
}
df = readFlie(input = 'C:\\Users\\zd200\\Desktop\\不常用\\直方图实例文件.txt',type = 'txt',row = F)
library(tidyr)
library(dplyr)
dt = df %>% pivot_longer(cols = -Group,names_to  = 'gene')
data1 = dt[,c(2,1,3)]
data2 = dt[,c(2,1,3)] %>% filter(Group!='Med')
p1 = PlotGroup_barplot(data = data1,test_method = 'kruskal.test',add = 'mean_sd')
p1
p2 = PlotGroup_barplot(data = data2,test_method = 'wilcox.test',add = 'mean_sd')
p2
savePlots(path = './PlotGroup_barplot1.pdf',plot = p1,type = 'pdf',width = 6,height = 4)
savePlots(path = './PlotGroup_barplot2.pdf',plot = p2,type = 'pdf',width = 6,height = 4)