Using an anonymous function within a function in R

Question

I have here a list of genetic loci containing their alleles encoded as three digit numbers, as class character. I have a few lines of code to go through the list and convert all instances to nucleic base letters (ie. A, C, G, T).

my_allele_list = list(loc1 = c("001", "002"),
                  loc2 = c("001", "003"),
                  loc3 = c("004", "001"),
                  loc4 = c("003", "003"),
                  loc5 = c("001", "002"),
                  loc6 = c("002", "004"))

a = c("001", "002", "003", "004")
b = c("A", "C", "G", "T")
for(i in seq_along(a)) my_allele_list <- 
  lapply(my_allele_list, function(x) gsub(a[i], b[i], x))

my_allele_list

So far so good, but to keep things tidy I would like to wrap these lines into a function.

convert_alleles <- function(x){
    a = c("001", "002", "003", "004")
    b = c("A", "C", "G", "T")
    for(i in seq_along(a)) x <- 
      lapply(x, function(x) gsub(a[i], b[i], x))
    }

convert_alleles(my_allele_list)

my_allele_list

However, as you can see on this second pass the function does not work - no error, just no change made to the list object. I suspect that the issue is with a clash with the anonymous function within the for loop. Can someone explain what the issue is and suggest a solution?

I have tried to change the variable name within the anonymous function and all references to it to y in case this was causing a conflict, but this has made no difference to the output. — rainbird, Dec 23 '21 at 20:33

akrun · Accepted Answer · 2021-12-23T20:41:53.917

It may be easier with a vectorized function as str_replace

library(dplyr)
library(purrr)
library(stringr)
map(my_allele_list, ~ str_replace_all(.x, setNames(b, a)))

-output

$loc1
[1] "A" "C"

$loc2
[1] "A" "G"

$loc3
[1] "T" "A"

$loc4
[1] "G" "G"

$loc5
[1] "A" "C"

$loc6
[1] "C" "T"

Also, if it is a fixed match and not a partial as in the example, use setNames to create a named vector and match and replace

map(my_allele_list, ~ unname(setNames(b, a)[.x]))
$loc1
[1] "A" "C"

$loc2
[1] "A" "G"

$loc3
[1] "T" "A"

$loc4
[1] "G" "G"

$loc5
[1] "A" "C"

$loc6
[1] "C" "T"

which can be also done with base R -lapply

lapply(my_allele_list, \(x) unname(setNames(b, a)[x]))
$loc1
[1] "A" "C"

$loc2
[1] "A" "G"

$loc3
[1] "T" "A"

$loc4
[1] "G" "G"

$loc5
[1] "A" "C"

$loc6
[1] "C" "T"

In the OP's function, the return value should be x

convert_alleles <- function(x){
     a = c("001", "002", "003", "004")
     b = c("A", "C", "G", "T")
     for(i in seq_along(a)) x <- 
       lapply(x, function(x) gsub(a[i], b[i], x))
     x
    }
convert_alleles(my_allele_list)
$loc1
[1] "A" "C"

$loc2
[1] "A" "G"

$loc3
[1] "T" "A"

$loc4
[1] "G" "G"

$loc5
[1] "A" "C"

$loc6
[1] "C" "T"

NOTE: when we run the function, it wouldn't change the object my_allele_list. For that we have assign back (<-)

my_allele_list <- convert_alleles(my_allele_list)

Thanks for the @akrun. Basic error, not printing the output so my bad. I'm rather new to function writing. But thanks also for suggesting some alternative approaches. — rainbird, Dec 23 '21 at 21:18
@rainbird in `R`, the output of last expression is returned. In your case, it is a `for` loop and thus it was not returned — akrun, Dec 23 '21 at 21:20

Using an anonymous function within a function in R

1 Answers1