Nested apply statement over a list of lists

user_7848253 0
Viewed 126

I want to pull out the proteins in seq.df (a single column df) that match the indices in the matching map.list (a list of lists).

Example data:

seq.df<- rbind.data.frame("MTHISPAVYGLWAIMSVLLAAFCAY",
    "MERSSAIVFPNVGTSVLSATIHLVGVTVLAHLISRRTALRGTST",
    "MLFEPFWCLLDLLRWSLDTHYIPAKRPLNGGGRSSNFD")
map.list<- list(a<- list(2,3,4,5,6,7),
    b<- list(13,14,30,31,32),
    c<- list(5,6,10,11))

Desired output: 

THISPA
GTAHL
PFLD

If I run a nested apply over just the first sublist of map.list, I get what I want for the first protein:

prot.list<- apply(seq.df, 1, function (x) lapply(map.list[[1]], function (y) substring(x, y, y)))

returns the expected result for the first sequence (THISPA,)

But I'm unsure how to get this function to iterate over all the sublists in map.list. I tried to wrap this into a for loop, but it does not give me the expected result:

for (i in seq_along(map.list)){
  each.map.list<- map.list[[i]]
  prot.list<- apply(seq.df, 1, function (x) lapply(each.map.list, function (y) substring(x, y, y)))
}

Output:

SPGL
SAPN
PFLD

I'd much rather add another lapply statement, but I'm not sure how to specify each list in map.list

#this does not work, but something like: 
prot.list<- apply(seq.df, 1, function (x) lapply(map.list, function (y) lapply([[y]], function (z) substring(x, z, z)))
lapply nested-loops r
3 Answers

We can use Map

unlist(Map(function(x, y) paste(substring(x, unlist(y), 
      unlist(y)), collapse=""), seq.df[[1]], map.list))
#[1] "THISPA" "GTAHL"  "PFLD"

Also, instead of unlisting twice, we can do a single unlist in the beginning and use that flattened list as input

l1 <- lapply(map.list, unlist)  
sapply(Map(substring, seq.df[[1]], first = l1, last = l1), paste, collapse="")
#[1] "THISPA" "GTAHL"  "PFLD"

Or with map2 from purrr

library(purrr)
map2_chr(seq.df[[1]], map.list, ~ str_c(substring(.x,
   unlist(.y), unlist(.y)), collapse=""))
user_37322710

Here is a solution using mapply()

It uses an anonymous function, using a character-split string of seq.df as x, and the list of positions as y.

mapply( function(x,y) paste0( x[ unlist(y) ], collapse = "" ), 
        x = stringr::str_split( seq.df[,1], pattern = ""),
        y = map.list )

[1] "THISPA" "GTAHL"  "PFLD"
user_63562780 
seq.df<- rbind.data.frame("MTHISPAVYGLWAIMSVLLAAFCAY",
                          "MERSSAIVFPNVGTSVLSATIHLVGVTVLAHLISRRTALRGTST",
                          "MLFEPFWCLLDLLRWSLDTHYIPAKRPLNGGGRSSNFD")
map.list<- list(a<- list(2,3,4,5,6,7),
                b<- list(13,14,30,31,32),
                c<- list(5,6,10,11))
lapply(1:nrow(seq.df), 
  function(x)paste(strsplit(as.character(seq.df[x,]), "")[[1]][unlist(map.list[[x]])], collapse=""))


[[1]]
[1] "THISPA"

[[2]]
[1] "GTAHL"

[[3]]
[1] "PFLD"
user_30879270
Related