将单列值连接到多个列名并扩展数据框

Question

I am creating a single summary table from multiple files. I have imported data from 4 files, file1...file4, and done some merging/manipulation using the reshape2 package, so my data looks like this:

 chr.list positions sample ref alt depth freq sum  min.prop
 chr1     12428     file4   C   a    52    2  14 0.2857143
 chr1     12428     file4   C   a    52    2  14 0.2857143
 chr1     12428     file3   C   c    52    1  18        NA
 chr1     12428     file3   C   g    52    2   4 0.5000000
 chr1     12428     file1   C   g    52    2   4 0.5000000
 chr1     12428     file2   C   t    52    2  16 0.1875000

Now, I want to separate the data for each of the four files but keep it in the same dataframe. I want to keep the chr.list , positions , ref and alt columns intact but want to remove the column sample , merge that column values with columns depth , freq , sum and min.prop and cast the data such that it looks like:

    chr.list    positions   ref alt file1.depth file1.freq  file1.sum   file1.min.prop  file2.depth file2.freq  file2.sum   file2.min.prop  file3.depth file3.freq  file3.sum   file3.min.prop
chr1    12428   C   a   NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  
chr1    12428   C   c   NA  NA  NA  NA  NA  NA  NA  NA  52  1   18  NA  
chr1    12428   C   g   52  2   4   0.5 NA  NA  NA  NA  52  2   4   0.5
chr1    12428   C   t   NA  NA  NA  NA  52  2   16  0.18    NA  NA  NA  NA

How can I do it? I am guessing using dcast but I am not sure.

Thanks!

Answer 1

The reshaping is straight-forward:

dd <- read.table(header = TRUE, stringsAsFactors = FALSE,
  text = "chr.list positions sample ref alt depth freq sum  min.prop
 chr1     12428     file4   C   a    52    2  14 0.2857143
 chr1     12428     file4   C   a    52    2  14 0.2857143
 chr1     12428     file3   C   c    52    1  18        NA
 chr1     12428     file3   C   g    52    2   4 0.5000000
 chr1     12428     file1   C   g    52    2   4 0.5000000
 chr1     12428     file2   C   t    52    2  16 0.1875000")

n <- names(dd)
rr <- reshape(dd[!duplicated(dd$sample), ], direction = 'wide', sep = '~',
              idvar = n[c(1:2,4:5)], v.names = n[6:9], timevar = n[3])

#   chr.list positions ref alt depth~file4 freq~file4 sum~file4 min.prop~file4
# 1     chr1     12428   C   a          52          2        14      0.2857143
# 3     chr1     12428   C   c          NA         NA        NA             NA
# 5     chr1     12428   C   g          NA         NA        NA             NA
# 6     chr1     12428   C   t          NA         NA        NA             NA
#   depth~file3 freq~file3 sum~file3 min.prop~file3 depth~file1 freq~file1
# 1          NA         NA        NA             NA          NA         NA
# 3          52          1        18             NA          NA         NA
# 5          NA         NA        NA             NA          52          2
# 6          NA         NA        NA             NA          NA         NA
#   sum~file1 min.prop~file1 depth~file2 freq~file2 sum~file2 min.prop~file2
# 1        NA             NA          NA         NA        NA             NA
# 3        NA             NA          NA         NA        NA             NA
# 5         4            0.5          NA         NA        NA             NA
# 6        NA             NA          52          2        16         0.1875

The order and column names isn't a reshape problem, so you need to do that yourself:

Find the variables you cast with the ~ , split by the tilde, reverse, and collapse back to the string. Then reorder the columns somehow

idx <- grepl('~', names(rr))
names(rr)[idx] <- sapply(strsplit(names(rr)[idx], '~'),
  function(x) paste0(rev(x), collapse = '_'))
rr[, c(1:4, order(names(rr)[-(1:4)]) + 4)]

#   chr.list positions ref alt file1_depth file1_freq file1_min.prop file1_sum
# 1     chr1     12428   C   a          NA         NA             NA        NA
# 3     chr1     12428   C   c          NA         NA             NA        NA
# 5     chr1     12428   C   g          52          2            0.5         4
# 6     chr1     12428   C   t          NA         NA             NA        NA
#   file2_depth file2_freq file2_min.prop file2_sum file3_depth file3_freq
# 1          NA         NA             NA        NA          NA         NA
# 3          NA         NA             NA        NA          52          1
# 5          NA         NA             NA        NA          NA         NA
# 6          52          2         0.1875        16          NA         NA
#   file3_min.prop file3_sum file4_depth file4_freq file4_min.prop file4_sum
# 1             NA        NA          52          2      0.2857143        14
# 3             NA        18          NA         NA             NA        NA
# 5             NA        NA          NA         NA             NA        NA
# 6             NA        NA          NA         NA             NA        NA