wig file

[prbndr]

hi all,

i'm trying to create a wig file in the following format:

track format=wiggle_0
variableStep chrom=2-micron span=1
1340 1
1341 1

where the left column represents a read start position and the right column represents the number of reads that start at that position. it is basically a frequency table that shows how many reads start at one position.

is there an easy way to create this kind of file from .bowtie or .sam/.bam output? any help would be appreciated!

[jujubix]

You could try using SAMtools to generate pileup output, from which you can use Unix cut on columns 2 and 4 to get read coverage per position (not number of reads).

If you feel a bit more adventurous and don't mind compiling C++, there is a small readCounter binary that takes in a bam file and outputs a wig file of reads starting at each position interval in the HMMcopy suite. Setting the interval to size 1 should hopefully get you the number of reads starting at each position.

[Richard Finney]

samtools view yourfile.bam | cut -f3,4 | \
awk ' { if($1!=c) { if(cnt){print $p" "cnt} print "variableStep chrom="$1 } else { if (p==$2){cnt++;} else{printf("%s %d\n",p,cnt);cnt=1;} } c=$1; p=$2; } END{print $p" "$cnt}'