如何将列表的列表转换为整齐的小标题或 R 中的 data.frame

How to convert list of list into tidy tibble or data.frame in R

我有以下列表:

my_lol <- structure(list(coolfactor_score = list(structure(c(0.164477631065473, 
0.198253819406019, 0.396414447052519, 0.133118603987442, 0.107735498488546
), .Names = c("B", "Mac", "NK", "Neu", "Stro")), structure(c(0.186215537135912, 
0.18408529174803, 0.375349920115798, 0.247664923324821, 0.006684327675438
), .Names = c("B", "Mac", "NK", "Neu", "Stro"))), sr_crt = list(
    structure(list(crt = 0.133118603987442, sr = 0.407076876403305), .Names = c("crt", 
    "sr")), structure(list(crt = 0.18408529174803, sr = 0.0829181742326453), .Names = c("crt", 
    "sr"))), sample_names = c("Sample1", "Sample2")), .Names = c("coolfactor_score", 
"sr_crt", "sample_names"))

看起来像这样:

> my_lol
$coolfactor_score
$coolfactor_score[[1]]
        B       Mac        NK       Neu      Stro 
0.1644776 0.1982538 0.3964144 0.1331186 0.1077355 

$coolfactor_score[[2]]
          B         Mac          NK         Neu        Stro 
0.186215537 0.184085292 0.375349920 0.247664923 0.006684328 


$sr_crt
$sr_crt[[1]]
$sr_crt[[1]]$crt
[1] 0.1331186

$sr_crt[[1]]$sr
[1] 0.4070769


$sr_crt[[2]]
$sr_crt[[2]]$crt
[1] 0.1840853

$sr_crt[[2]]$sr
[1] 0.08291817



$sample_names
[1] "Sample1" "Sample2"
# Note that the number of samples can be more than 2 and cell type more than 5.

如何将其整理到此数据框中(小标题)

CellType    Sample    CoolFactorScore  SR            CRT
B           Sample1   0.1644776        0.4070769     0.1331186
Mac         Sample1   0.1982538        0.4070769     0.1331186
NK          Sample1   0.3964144        0.4070769     0.1331186
Neu         Sample1   0.1331186        0.4070769     0.1331186
Stro        Sample1   0.1077355        0.4070769     0.1331186
B           Sample2   0.186215537      0.08291817    0.1840853
Mac         Sample2   0.184085292      0.08291817    0.1840853
NK          Sample2   0.375349920      0.08291817    0.1840853
Neu         Sample2   0.247664923      0.08291817    0.1840853
Stro        Sample2   0.006684328      0.08291817    0.1840853

一种使用基础 R 的方法:

mylist <- lapply(1:2, function(i) {
  #this is the important bit where you extract the corresponding elements
  #of sample 1 first and sample 2 second.
  df <- data.frame(lapply(my_lol, '[', i))
  names(df) <- c('CoolFactorScore', 'CRT', 'SR', 'Sample')
  df$CellType <- rownames(df)
  row.names(df) <- NULL
  df
})

do.call(rbind, mylist)

输出:

  CoolFactorScore       CRT         SR  Sample CellType
1      0.164477631 0.1331186 0.40707688 Sample1        B
2      0.198253819 0.1331186 0.40707688 Sample1      Mac
3      0.396414447 0.1331186 0.40707688 Sample1       NK
4      0.133118604 0.1331186 0.40707688 Sample1      Neu
5      0.107735498 0.1331186 0.40707688 Sample1     Stro
6      0.186215537 0.1840853 0.08291817 Sample2        B
7      0.184085292 0.1840853 0.08291817 Sample2      Mac
8      0.375349920 0.1840853 0.08291817 Sample2       NK
9      0.247664923 0.1840853 0.08291817 Sample2      Neu
10     0.006684328 0.1840853 0.08291817 Sample2     Stro

这里有一个不太优雅的方法:

int <- lapply(1:2, function(x) do.call(data.frame, 
              c(list(CoolFactorScore=my_lol[[1]][[x]]), 
                my_lol[[2]][[x]], 
                list(Sample=my_lol[[3]][[x]])))) 
do.call(rbind, int)

      CoolFactorScore       crt         sr  Sample
B         0.164477631 0.1331186 0.40707688 Sample1
Mac       0.198253819 0.1331186 0.40707688 Sample1
NK        0.396414447 0.1331186 0.40707688 Sample1
Neu       0.133118604 0.1331186 0.40707688 Sample1
Stro      0.107735498 0.1331186 0.40707688 Sample1
B1        0.186215537 0.1840853 0.08291817 Sample2
Mac1      0.184085292 0.1840853 0.08291817 Sample2
NK1       0.375349920 0.1840853 0.08291817 Sample2
Neu1      0.247664923 0.1840853 0.08291817 Sample2
Stro1     0.006684328 0.1840853 0.08291817 Sample2

这是一个没有循环的解决方案,使用了 data.table 包的功能。

library(data.table)

第 1 步:展开列表

unlist(my_lol) -> tmp1

第 2 步:转置并转换为 data.table
这样您将获得可以由原始数据组成的最宽的 table 。它应该根据要求(在进一步的步骤中)转换为 long table。

as.data.table(t(tmp1)) -> tmp2

第三步:需要将'sample_names1'和'sample_names2'手动转换为'Sample'。
如果您想泛化到多个 sample_names 值,那么您应该根据可能值的语法修改此步骤。(此版本适用于此类 sample_names 值语法如:'Sample1'、'Sample2'、'Sample3' 等等。)

names(tmp2) <- gsub('sample_names\d+', 'Sample', names(tmp2))

第四步:根据tmp2的字段名创建度量字段名table

measure <- unique(names(tmp2))

第 5 步:从宽 table (tmp2)

创建更长的 table (tmp3)
tmp3 <- melt(tmp2, 
             measure.vars = patterns(measure), 
             value.name = measure)

第 6 步:根据要求重命名列

names(tmp3) <- gsub('coolfactor_score.', '', names(tmp3))
names(tmp3) <- gsub('sr_crt.', '', names(tmp3))
setnames(tmp3, 'crt', 'CRT')
setnames(tmp3, 'sr', 'SR')

第 7 步:从 tmp3

创建更长的 table (mylist)
mylist <- melt(tmp3,
               id.vars = c('Sample',
                           'CRT',
                           'SR'),
               measure.vars = c('B', 
                                'Mac',  
                                'NK',   
                                'Neu',
                                'Stro'),
               value.name = 'CoolFactorScore',
               variable.name = 'CellType')

第 8 步:根据要求对列重新排序

setcolorder(mylist, c('CellType', 'Sample', 'CoolFactorScore', 'SR', 'CRT'))

第 9 步:根据请求重新排序行

mylist <- mylist[order(Sample, CellType)]