Converting letters to numbers in entire dataframe

Question

I am having trouble applying the chartr() function on all columns of my data frame for converting letters to numbers.

I managed doing it on single colum

Answer 1

Using dplyr

library(dplyr)
DF %>% mutate_each(funs(chartr("ACGT", "1234", .)))

You get:

  ID POS1 POS2 POS3
1  1   13   34   33
2  2   32   22   24
3  3   44   42   14

Answer 2

> cbind(DF, setNames( lapply( DF[-1], chartr, old='ACGT', new='1234'), 
                      paste0("POS", 1:(length(DF)-1),"X"))  )
  ID POS1 POS2 POS3 POS1X POS2X POS3X
1  1   AG   GT   GG    13    34    33
2  2   GC   CC   CT    32    22    24
3  3   TT   TC   AT    44    42    14

Answer 3

Why not use lapply?

DF2 <- DF ## to not overwrite the original DF
DF2[-1] <- lapply(DF2[-1], chartr, old = "ACGT", new = "1234")
DF2
#   ID POS1 POS2 POS3
# 1  1   13   34   33
# 2  2   32   22   24
# 3  3   44   42   14

Now you have two data frames with equivalent column names which I find easier to compare than appending new columns to the old data. Especially when there are many many columns.

Answer 4

You could also use mgsub from qdap

library(qdap)
DF[paste0('POS', 1:3,'X')] <- mgsub(c('A', 'C', 'G', 'T'), 1:4, 
               as.matrix(DF[-1]))
DF
#  ID POS1 POS2 POS3 POS1X POS2X POS3X
#1  1   AG   GT   GG    13    34    33
#2  2   GC   CC   CT    32    22    24
#3  3   TT   TC   AT    44    42    14