如何在多个数据帧中找到通用和唯一的元素

Question

Objective来查看常见的元素，即我的行，它们基本上是我的不同比较中的基因名称。

这就是我试图遵循的answer。

df1 = data.frame(genes = c('gene1', 'gene3', 'gene4', 'gene2'))
df2 = data.frame(genes = c('gene3', 'gene2', 'gene5', 'gene1', "genet"))
df3 = data.frame(genes = c('gene6', 'gene3', 'gene4', 'gdene7', 'genex', "gene10"))
dfList <- list(df1, df2, df3) 

reduce(dfList, inner_join)
reduce(dfList, inner_join)
Joining, by = "genes"
Joining, by = "genes"
  genes
1 gene3

这在这种情况下失败了。

    df1 = data.frame(genes = c('gene1', 'gene3', 'gene4', 'gene2'))
    df2 = data.frame(genes = c('gene3', 'gene2', 'gene5', 'gene1', "genet"))
    df3 = data.frame(genes = c('gene6', 'gene13', 'gene4', 'gdene7', 'genex', "gene10"))
    dfList <- list(df1, df2, df3) 
    
    reduce(dfList, inner_join)

educe(dfList, inner_join)
Joining, by = "genes"
Joining, by = "genes"
[1] genes
<0 rows> (or 0-length row.names)

现在如何解决这个问题。我给了一个小的，我有15个比较。

Expected output 
  gene3   df1 df2 df3 ## for common genes

gene1  df1 df2  ## for genes which arr not across all the combination 
gene2  

在第一种情况下，解决方案在所有情况下都是预先设置的gene3，但在只有2种情况下出现时失败。

那么，我如何找出所有可能的组合，当基因在不同的可能组合中存在时。

例如，如果、gene3、都存在，那么它是报告的，但是在df1和df2中存在gene1和gene2，但是没有报告这些。

因此，我想看看一组基因是否存在于所有条件下，这些基因最不可能存在，而是存在于其存在的所有可能的组合中。

我的实际数据文件就是这样命名的，它位于列表中。

names(result_abd)
 [1] "M0_vs_M1_TCGA_stages" "M0_vs_M2_TCGA_stages" "M0_vs_M3_TCGA_stages" "M0_vs_M4_TCGA_stages" "M0_vs_M5_TCGA_stages" "M1_vs_M2_TCGA_stages"
 [7] "M1_vs_M3_TCGA_stages" "M1_vs_M4_TCGA_stages" "M1_vs_M5_TCGA_stages" "M2_vs_M3_TCGA_stages" "M2_vs_M4_TCGA_stages" "M2_vs_M5_TCGA_stages"
[13] "M3_vs_M4_TCGA_stages" "M3_vs_M5_TCGA_stages" "M4_vs_M5_TCGA_stages"
> 

所以我希望每一次数据都有15列

我运行了您的代码，输出如下

dput(head(a))
structure(list(gene = c("ENSG00000000003", "ENSG00000000971", 
"ENSG00000002726", "ENSG00000003989", "ENSG00000005381", "ENSG00000006534"
), dfM0_vs_M1_TCGA_stages = c("M0_vs_M1_TCGA_stages", "M0_vs_M1_TCGA_stages", 
"M0_vs_M1_TCGA_stages", "M0_vs_M1_TCGA_stages", "M0_vs_M1_TCGA_stages", 
"M0_vs_M1_TCGA_stages"), dfM0_vs_M2_TCGA_stages = c(NA, "M0_vs_M2_TCGA_stages", 
"M0_vs_M2_TCGA_stages", NA, "M0_vs_M2_TCGA_stages", NA), dfM0_vs_M3_TCGA_stages = c("M0_vs_M3_TCGA_stages", 
"M0_vs_M3_TCGA_stages", "M0_vs_M3_TCGA_stages", NA, "M0_vs_M3_TCGA_stages", 
NA), dfM0_vs_M4_TCGA_stages = c("M0_vs_M4_TCGA_stages", NA, "M0_vs_M4_TCGA_stages", 
NA, "M0_vs_M4_TCGA_stages", "M0_vs_M4_TCGA_stages"), dfM0_vs_M5_TCGA_stages = c("M0_vs_M5_TCGA_stages", 
NA, "M0_vs_M5_TCGA_stages", NA, "M0_vs_M5_TCGA_stages", "M0_vs_M5_TCGA_stages"
), dfM1_vs_M2_TCGA_stages = c(NA_character_, NA_character_, NA_character_, 
NA_character_, NA_character_, NA_character_), dfM1_vs_M3_TCGA_stages = c(NA, 
NA, NA, NA, "M1_vs_M3_TCGA_stages", NA), dfM1_vs_M4_TCGA_stages = c(NA, 
"M1_vs_M4_TCGA_stages", NA, NA, NA, NA), dfM1_vs_M5_TCGA_stages = c(NA, 
NA, "M1_vs_M5_TCGA_stages", NA, NA, NA), dfM2_vs_M3_TCGA_stages = c(NA, 
NA, NA, NA, "M2_vs_M3_TCGA_stages", NA), dfM2_vs_M4_TCGA_stages = c(NA, 
"M2_vs_M4_TCGA_stages", NA, NA, NA, NA), dfM2_vs_M5_TCGA_stages = c(NA, 
NA, "M2_vs_M5_TCGA_stages", NA, "M2_vs_M5_TCGA_stages", NA), 
    dfM3_vs_M4_TCGA_stages = c(NA, "M3_vs_M4_TCGA_stages", NA, 
    NA, "M3_vs_M4_TCGA_stages", NA), dfM3_vs_M5_TCGA_stages = c(NA, 
    "M3_vs_M5_TCGA_stages", NA, NA, "M3_vs_M5_TCGA_stages", NA
    ), dfM4_vs_M5_TCGA_stages = c(NA, NA, "M4_vs_M5_TCGA_stages", 
    NA, NA, NA)), row.names = c(NA, -6L), class = c("tbl_df", 
"tbl", "data.frame"))

Dataframe格式

 A tibble: 6 × 16
  gene            dfM0_vs_M1_TCGA… dfM0_vs_M2_TCGA… dfM0_vs_M3_TCGA… dfM0_vs_M4_TCGA… dfM0_vs_M5_TCGA… dfM1_vs_M2_TCGA… dfM1_vs_M3_TCGA… dfM1_vs_M4_TCGA…
  <chr>           <chr>            <chr>            <chr>            <chr>            <chr>            <chr>            <chr>            <chr>           
1 ENSG00000000003 M0_vs_M1_TCGA_s… NA               M0_vs_M3_TCGA_s… M0_vs_M4_TCGA_s… M0_vs_M5_TCGA_s… NA               NA               NA              
2 ENSG00000000971 M0_vs_M1_TCGA_s… M0_vs_M2_TCGA_s… M0_vs_M3_TCGA_s… NA               NA               NA               NA               M1_vs_M4_TCGA_s…
3 ENSG00000002726 M0_vs_M1_TCGA_s… M0_vs_M2_TCGA_s… M0_vs_M3_TCGA_s… M0_vs_M4_TCGA_s… M0_vs_M5_TCGA_s… NA               NA               NA              
4 ENSG00000003989 M0_vs_M1_TCGA_s… NA               NA               NA               NA               NA               NA               NA              
5 ENSG00000005381 M0_vs_M1_TCGA_s… M0_vs_M2_TCGA_s… M0_vs_M3_TCGA_s… M0_vs_M4_TCGA_s… M0_vs_M5_TCGA_s… NA               M1_vs_M3_TCGA_s… NA              
6 ENSG00000006534 M0_vs_M1_TCGA_s… NA               NA               M0_vs_M4_TCGA_s… M0_vs_M5_TCGA_s… NA               NA               NA   

这就是我想要的。下一步作为我想要看到的例子

如果我认为这种基因ENSG00000000971存在于7种比较中，而在其他被报道为NA.How的地方则不存在，我就将它们分组。

就像用这些基因构建另一个数据框架一样，我们可以说，在多重比较中，这些基因是存在的，而不包括在这是NA的地方。

Answer 1

我不清楚您希望如何格式化您的文件(几个索引列或一个包含字符串或列表列的列)。但这里有一个选择。首先，我将您的数据名列表合并到一个数据框架中，并有一个指示源的索引。

library(tidyverse)

dfList <- list(df1, df2, df3) 

dfList %>% 
  bind_rows(.id="df") %>% 
  pivot_wider(names_from=df, names_prefix="df", values_from=df) 
# A tibble: 10 × 4
   genes  df1   df2   df3  
   <chr>  <chr> <chr> <chr>
 1 gene1  1     2     NA   
 2 gene3  1     2     3    
 3 gene4  1     NA    3    
 4 gene2  1     2     NA   
 5 gene5  NA    2     NA   
 6 genet  NA    2     NA   
 7 gene6  NA    NA    3    
 8 gdene7 NA    NA    3    
 9 genex  NA    NA    3    
10 gene10 NA    NA    3    

以下是对执行部分问题的补充。(不过请注意，这实际上是一个新问题，应该是一个新帖子。)

dfList %>% 
  bind_rows(.id="df") %>% 
  group_by(genes) %>% 
  summarise(minDF=min(df), maxDF=max(df)) %>% 
  filter(minDF == maxDF & maxDF == 3) %>% 
  pull(genes)
[1] "gdene7" "gene10" "gene6"  "genex" 

再一次，关键是将所有数据放到一个数据帧中。(所需的输出格式不清楚。)