Hey! I have a set of sequences that I got from differential ChIP-seq analysis. I would like to use MEME-ChIP for motif discovery on this set of sequences, as input it requires sequences of the same length, however, for this trimming I would like to take the 400bp on the center of each of the sequences from my list of different length sequences (ranging from 450 to 1400 approx). Could anybody help me figure out how can I do this in Biostrings or some other R package? Thanks a lot!

Biostrings allows you to subset stings using the index of the characters (or bp). Here is a function that finds center n characters of a string. I've added a "wobble" for cases when the exact center doesn't exist (selecting an odd amount of characters out out an even length string and vice-versa). Hope this helps :) -Louise

    library(Biostrings)

get_center <- function(bstr, center_n){
  outer_n = length(bstr) - center_n
  if(outer_n < 0) stop( "Center is longer than string")

  half_outer_n <- outer_n %/% 2
  wobble = 0
  if (outer_n %% 2 == 1){  
    # if remainder cannot be split evenly 'wobble' start index
    wobble = sample(c(0,1), 1)
  }
  start_i <- half_outer_n + 1 + wobble
  end_i <- half_outer_n + center_n + wobble
  return(bstr[start_i:end_i])
}

long_bs <- BString("aaaaaaaaaatttttttttcttttttttttgggggggggg")
short_bs_even <- BString("attg")
short_bs_odd <- BString("atctg")

get_center(long_bs, 20)
# 20-letter BString object
# seq: tttttttttctttttttttt

get_center(short_bs_even, 2)
# 2-letter BString object
# seq: tt
get_center(short_bs_odd, 2)
# 2-letter BString object
# seq: ct
# OR
# 2-letter BString object
# seq: tc
get_center(ss, 5)
# Error in get_center(ss, 5) : Center is longer than string