「数据整理」使用tidyr包和dplyr包进行数据清洗

最近处理扩增子数据，很多数据需要转化整理，从而适用于ggplot2系列的可视化工作，之前常常在Excel表格中手动整理，今天记录写一下使用tidyr包和dplyr包进行数据整理的过程，最后使用ggplot2进行批量作图，cowplot包对图进行组合。

1.载入包

library(tidyverse)
list.files()

---

2.长宽数据转换

family_data <- read_tsv('C:/Users/Administrator/Documents/R_work/03_BD_L_microbiome/00_rawdata/outfiles/expr.relative_abundance.abfam.txt')
head(family_data)
# A tibble: 6 x 19
  Family `Bd-1-1` `Bd-1-2` `Bd-1-3` `Bd-1-4` `Bd-1-5` `Bd-1-6` `Bd-2-1` `Bd-2-2` `Bd-2-3`
  <chr>     <dbl>    <dbl>    <dbl>    <dbl>    <dbl>    <dbl>    <dbl>    <dbl>    <dbl>
1 Aceto~ 0.0224    0.0337   0.0398   5.63e-2   0.114   7.77e-2 0.00396    0.0648  2.23e-2
2 Acida~ 0.0109    0.00232  0.00648  4.22e-5   0       0.      0          0       1.61e-3
3 Actin~ 0.000613  0        0        0.        0       3.68e-5 0          0       0.     
4 Akker~ 0         0        0        0.        0       0.      0.000967   0       0.     
5 Archa~ 0         0        0        0.        0       0.      0.000193   0       8.79e-5
6 Bacil~ 0.0295    0.0549   0.0439   7.11e-2   0.0383  6.37e-2 0.0343     0.0153  1.64e-2
# ... with 9 more variables: `Bd-2-4` <dbl>, `Bd-2-5` <dbl>, `Bd-2-6` <dbl>, `Bd-3-1` <dbl>,
#   `Bd-3-2` <dbl>, `Bd-3-3` <dbl>, `Bd-3-4` <dbl>, `Bd-3-5` <dbl>, `Bd-3-6` <dbl>

#宽数据转为长数据
family_data <- family_data %>%
  pivot_longer(!Family, names_to = "Sample", values_to = "value")

head(family_data)
# A tibble: 1,278 x 3
   Family           Sample   value
   <chr>            <chr>    <dbl>
 1 Acetobacteraceae Bd-1-1 0.0224 
 2 Acetobacteraceae Bd-1-2 0.0337 
 3 Acetobacteraceae Bd-1-3 0.0398 
 4 Acetobacteraceae Bd-1-4 0.0563 
 5 Acetobacteraceae Bd-1-5 0.114  
 6 Acetobacteraceae Bd-1-6 0.0777 
 7 Acetobacteraceae Bd-2-1 0.00396
 8 Acetobacteraceae Bd-2-2 0.0648 
 9 Acetobacteraceae Bd-2-3 0.0223 
10 Acetobacteraceae Bd-2-4 0.00813
# ... with 1,268 more rows

---

3.根据Sample列生成新列

family_data <- family_data %>% 
  mutate(Group = case_when(
    startsWith(Sample, "Bd-1") ~ "First ",
    startsWith(Sample, "Bd-2") ~ "Second ",
    startsWith(Sample, "Bd-3") ~ "Third "),
    num = case_when(
      startsWith(Sample, "Bd-1") ~ 1,
      startsWith(Sample, "Bd-2") ~ 2,
      startsWith(Sample, "Bd-3") ~ 3))

head(family_data)
# A tibble: 1,278 x 5
   Family           Sample   value Group           num
   <chr>            <chr>    <dbl> <chr>         <dbl>
 1 Acetobacteraceae Bd-1-1 0.0224  First       1
 2 Acetobacteraceae Bd-1-2 0.0337  First       1
 3 Acetobacteraceae Bd-1-3 0.0398  First       1
 4 Acetobacteraceae Bd-1-4 0.0563  First       1
 5 Acetobacteraceae Bd-1-5 0.114   First       1
 6 Acetobacteraceae Bd-1-6 0.0777  First       1
 7 Acetobacteraceae Bd-2-1 0.00396 Second      2
 8 Acetobacteraceae Bd-2-2 0.0648  Second      2
 9 Acetobacteraceae Bd-2-3 0.0223  Second      2
10 Acetobacteraceae Bd-2-4 0.00813 Second      2
# ... with 1,268 more rows

---

4.批量作图

#写一个作图函数
myboxplot.v1 <-  function(df, taxo) {
  p <- ggplot() + 
    geom_boxplot(data = filter(df, Family == taxo),
                 aes(x = Group, y = value*100, fill = Group)) +
    geom_smooth(data = filter(df, Family  == taxo),
                aes(x = num, y = value*100),
                size = 1.2, color = "grey40",level = 0.8, alpha = 0.3) +
    labs(y = "Relative abundance (%)", x = "") + 
    ggtitle(taxo) +
    theme(plot.title = element_text(size = 14, hjust = 0.5, face = "bold"), 
          strip.text = element_text(size = rel(3)),
          axis.title.x = element_text(size = 14, vjust = 0.5, 
                                      hjust = 0.5),
          axis.title.y = element_text(size = 13, vjust = 0.5, 
                                      hjust = 0.5,face = "bold",color ="black"),
          axis.text.x = element_text(angle = 0, size = 9,
                                     vjust = 0.5, hjust = 0.5,
                                     face = "bold",color = "black"),
          axis.text.y = element_text(size = 12,vjust = 0.5, 
                                     hjust = 0.5),
          axis.line = element_line(colour = "black"),
          panel.background = element_rect(fill = NA)) + 
    scale_fill_manual(values = c("#2874C5", "#EABF00", "#E64B35B2")) +
    scale_x_discrete(labels = c('1st','2nd','3rd')) +
    guides(fill = FALSE) 
}

#读入需要作图的科水平菌群名
#将相对丰度大于1%的科水平菌群分为下降趋势类群和上升趋势类群，分别作图,并添加拟合曲线
down <- read.table("C:/Users/Administrator/Documents/R_work/03_BD_L_microbiome/00_rawdata/outfiles/02_down_family_ID.txt",as.is = T,row.names = 1 )
row.names(down)
[1] "Streptococcaceae" "Rhizobiaceae"     "Acetobacteraceae" "Bacillaceae"     
[5] "Burkholderiaceae" "Pseudomonadaceae"
up <- read.table("C:/Users/Administrator/Documents/R_work/03_BD_L_microbiome/00_rawdata/outfiles/03_up_family_ID.txt", as.is = T, row.names = 1)
row.names(up)
[1] "Enterobacteriaceae" "Holosporaceae"      "Leuconostocaceae"   "Lactobacillaceae" 

#批量作图，并拼图
list_box1 <- lapply(row.names(up), myboxplot.v1, df = family_data)
mult_plot1 <- plot_grid(plotlist=list_box1, ncol= 2, nrow = 2)

---

Rplot04.jpeg

list_box2 <- lapply(row.names(down), myboxplot.v1, df = family_data)
mult_plot2 <- plot_grid(plotlist=list_box2, ncol= 3, nrow = 2)

---

Rplot03.jpeg

这个学习笔记记录了tidyr包pivot_longer函数将宽数据转化为长数据，dplyr包mutate函数生成新的数据列，最后学习使用function函数将ggplot2画图代码封装成一个函数，配合lapply函数对数据进行批量作图，最后使用cowplot包的plot_grid函数进行多图组合。