Hi I have a text file that I want to change some of the characters based on their position. My file contain multiple lines and characters should be counted continuously line by line. For example, I want to convert the 150th T to C. What can I do? Here is a portion of my file:

GATCACAGGTCTATCACCCTATTAACCACTCACGGGAGCTCTCCATGCAT
TTGGTATTTTCGTCTGGGGGGTGTGCACGCGATAGCATTGCGAGACGCTG
GAGCCGGAGCACCCTATGTCGCAGTATCTGTCTTTGATTCCTGCCTCATT
CTATTATTTATCGCACCTACGTTCAATATTACAGGCGAACATACCTACTA
AAGTGTGTTAATTAATTAATGCTTGTAGGACATAATAATAACAATTGAAT
GTCTGCACAGCCGCTTTCCACACAGACATCATAACAAAAAATTTCCACCA
AACCCCCCCCTCCCCCCGCTTCTGGCCACAGCACTTAAACACATCTCTGC
CAAACCCCAAAAACAAAGAACCCTAACACCAGCCTAACCAGATTTCAAAT
TTTATCTTTAGGCGGTATGCACTTTTAACAGTCACCCCCCAACTAACACA
TTATTTTCCCCTCCCACTCCCATACTACTAATCTCATCAATACAACCCCC
GCCCATCCTACCCAGCACACACACACCGCTGCTAACCCCATACCCCGAAC
CAACCAAACCCCAAAGACACCCCCCACAGTTTATGTAGCTTACCTCCTCA
AAGCAATACACTGAAAATGTTTAGACGGGCTCACATCACCCCATAAACAA
ATAGGTTTGGTCCTAGCCTTTCTATTAGCTCTTAGTAAGATTACACATGC
AAGCATCCCCGTTCCAGTGAGTTCACCCTCTAAATCACCACGATCAAAAG
GGACAAGCATCAAGCACGCAGCAATGCAGCTCAAAACGCTTAGCCTAGCC
ACACCCCCACGGGAAACAGCAGTGATTAACCTTTAGCAATAAACGAAAGT
TTAACTAAGCTATACTAACCCCAGGGTTGGTCAATTTCGTGCCAGCCACC
GCGGTCACACGATTAACCCAAGTCAATAGAAGCCGGCGTAAAGAGTGTTT

try this...

awk -F "" '{if((max+NF)>150){for(i=1;i<=NF;i++){if((max+i) == 150 && $i ~ /T/){$i = "C"}}}else{max+=NF}}1' file

awk -v p=150 -v l="C" '{for (i=1; i<=length($0); i++) {++c; o=$0; if (c==p) o=substr($0,1,i-1) l substr($0,i+1);};print o;}' infile

Here is a much longer alternative way to do this:

#!/bin/ksh
# Usage: tester [count [from [to]]]
#       Change the "count"th occurrence of the character specified by "from"
#       to the character specified by "to".  If not given on the command line,
#               "count" defaults to 150,
#               "from" defaults to "T", and
#               "to" defaujlts to "C".
awk -v cnt=${1:-150} -v from=${2:-T} -v to=${3:-C} 'BEGIN{tmpc = "\a"}
cnt>0 {
        # See if changing every "from" character on this line will go too far.
        if((n = gsub(from, from)) < cnt) {
                # No.  Reduce cnt by the number of "from" characters found and
                # print the unchanged line.
                cnt -= n
        } else {
                # The "from" character we need to change is on this line.
                # Change cnt - 1 "from" characters to "tmpc" characters.
                for(i = 1; i < cnt; i++) sub(from, tmpc)
                # Change the desired "from" charaacter to the "to" character.
                sub(from, to)
                # Change the "tmpc" characters inserted above back to "from"
                # characters.
                gsub(tmpc, from)
                # Note that we are done looking.
                cnt = 0
        }
        # Fall through to next action to print the processed line.
}
1
END {   if(cnt) {
                printf("Still looking for %d %s characters when EOF found.\n",
                        cnt, from)
                exit 1
        }
        exit 0
}' in

Hi Guys,
Thanks for the codes you sent. However, they don't work properly when I change the the number of the letter. Here I've attached the file which may help you work on it a little bit more easily. Thank you agian for putting effort and time on my problem.

Attached Files

mtDNA_GATK_reference_letters.txt (16.8 KB, 5 views)

Which letter do you want to change?
Which letter do you want to replace it?
Which occurrence of that latter do you want to change?

You said my script doesn't work properly. What does that mean? Did it produce a diagnostic message? Was some other character in the file changed? What system are you using?

When I try to change the 150th occurrence of T to C in the file you attached using the script I provided, the T in column 20 of line 13 does change from T to C just as you said you wanted.

Looking more closely at your input file I see that there are 4998 As, 5095 Cs, 2108 Gs, 3997 Ts, and 332 newlines in a file that contains 17,235 bytes. Could the problem be that there are four sequences of lowercase letters in the big file you attached, but only uppercase letters in the sample that you said was representative of your entire file?

awk -v p=150 -v r="T" -v l="C" '{for (i=1; i<=length($0); i++) {s=substr($0,i,1);if (s~r)n++;if (n==p)s=l;printf s}print ""}' infile