R - convert a string of repeat amino acids from e.g. NNNN to (N4)

user_6539898 0
Viewed 446

I am computing haplotypes from multiple sequence alignments and am getting stretches of repeats such as RNNNNNNNT and RNNNT. There are many variations, making it hard to make sense of the data.

The data is listed as below and am I'm interested in generating the column haplotypes_2 based on haplotypes_1 as indicated:

hap_code    haplotypes_1          haplotypes_2
  1    SKNNNRNNNNNKNNNNNNNKF    SK(N3)R(N5)K(N7)KF
  2    SKNNNNNNNNNKNNNNNNNNKF   SK(N9)K(N8)KF
  3    SKNNNNNNNNNNNNNNNNKF     SK(N16)KF
r regex
6 Answers

Using stringr and a custom function:

library(stringr)

replace_string <- function(x) {
  sprintf("(%s%i)", str_sub(x, end = 1L), str_length(x))
}

df1$hapnew <- str_replace_all(df1$haplotypes_1, "N+", replace_string)

  hap_code           haplotypes_1       haplotypes_2             hapnew
1        1  SKNNNRNNNNNKNNNNNNNKF SK(N3)R(N5)K(N7)KF SK(N3)R(N5)K(N7)KF
2        2 SKNNNNNNNNNKNNNNNNNNKF      SK(N9)K(N8)KF      SK(N9)K(N8)KF
3        3   SKNNNNNNNNNNNNNNNNKF          SK(N16)KF          SK(N16)KF
user_32421300 
x = c("SKNNNRNNNNNKNNNNNNNKF", "SKNNNNNNNNNKNNNNNNNNKF", "SKNNNNNNNNNNNNNNNNKF")
sapply(strsplit(x, ""), function(mystr)
    with(rle(mystr),
         paste(paste0(ifelse(lengths > 1, paste0("(",values), values),
                      ifelse(lengths > 1, paste0(lengths,")"), ""),
                      collapse = ""))))
#[1] "SK(N3)R(N5)K(N7)KF" "SK(N9)K(N8)KF"      "SK(N16)KF"
user_71289340

Here is an option using gsubfn and str_count

library(stringr)
library(gsubfn)
gsubfn("N+", ~paste0("(", substr(x, 1, 1), str_count(x), ")"), df1$haplotypes_1)
#[1] "SK(N3)R(N5)K(N7)KF" "SK(N9)K(N8)KF"      "SK(N16)KF"

Or as @G.Grothendieck suggested, the str_count can be replaced with nchar

gsubfn("N+", ~sprintf("(%0.1s%d)", x, nchar(x)), df1$haplotypes_1) 
#[1] "SK(N3)R(N5)K(N7)KF" "SK(N9)K(N8)KF"      "SK(N16)KF"
user_37322710 
prev <- ""
count <- 1
output <- ""
for (character in string) {
  if (character==prev) {
    count <- count + 1
  }
  else {
    if (count > 1) {
      output <- output + prev + toString(count)
    }
    else {
      output <- output + prev
    }
  }
  prev <- character
}

This alludes to my comment, there may be a problem hiding in this but the gist is there.

user_13916250

I will suggest another one: 

df <- data.frame(haplotypes_1 = c("SKNNNRNNNNNKNNNNNNNKF", 
                                  "SKNNNNNNNNNKNNNNNNNNKF", 
                                  "SKNNNNNNNNNNNNNNNNKF"))
df$haplotypes_2 <- sapply(df$haplotypes_1, function(q){
  x <- rle(strsplit(as.character(q), "N")[[1]])
  res <- rep("", length(x$lengths))
  res[x$lengths != 1] <- paste("(N", x$lengths[x$lengths != 1], ")", sep="")
  res[x$lengths == 1] <- x$values[x$values != ""]
  paste(res, collapse = "")
})
user_13766520

Almost exactly the same as @d.b, but converted to a couple functions so it's reusable and reader-friendly:

abbreviate_letters <- function(type_letters) {
  runs <- rle(type_letters)
  run_codes <- ifelse(
    runs[["lengths"]] == 1,
    yes = runs[["values"]],
    no  = paste0("(", runs[["values"]], runs[["lengths"]], ")")
  )
  paste0(run_codes, collapse = "")
}


condense_haplotype <- function(haplotype_long) {
  split_terms <- strsplit(haplotype_long, split = "")
  vapply(
    X         = split_terms,
    FUN       = abbreviate_letters,
    FUN.VALUE = character(1)
  )
}


haplotypes <- c(
  "SKNNNRNNNNNKNNNNNNNKF",
  "SKNNNNNNNNNKNNNNNNNNKF",
  "SKNNNNNNNNNNNNNNNNKF"
)

condense_haplotype(haplotypes)
# [1] "SK(N3)R(N5)K(N7)KF" "SK(N9)K(N8)KF"      "SK(N16)KF"
user_49122290
Related