how to keep column 6 (normalized tag count) in peaks.txt file called by Homer callpeaks after pos2bed manipulation?

I use HOMER to call peaks getting peaks.txt file. Then I use pos2bed.pl to transform peaks.txt to peaks.bed However, the column 6 loss after the transform, which showed the normalized tag count (equal to RPKM reflecting peak density).

chip-seq

• 4,275 views • link

updated 8.8 years ago by Prakash

• written 8.8 years ago by Ming Lu •

2 answers

simple "grep" and "awk" can do your job.

grep -v "#" peak.txt |cut -f 1,2,3,4,6 | awk '{print $2"\t"$3"\t"$4"\t"$1"\t"$5}' >peak.bed

• 0 views • link

written 8.8 years ago by Prakash

Thank you, I use these code, and the column 6 will be kept after bedtools intersect

cut -f 1,2,3,4,6 peaks.txt | awk '{print $2"\t"$3"\t"$4"\t"$1"\t"$5}' >peak1.bed
pos2bed.pl peaks.txt > peak2.bed
awk 'NR==FNR {h[$4] = $5; next} {print $1"\t"$2"\t"$3"\t"$4"\t"$5"\t"$6"\t"h[$4]}' peak1.bed peak2.bed >peaks.bed
chr11   117467921   117468098   chr11-2 1   +   86.8
chr17   39636555    39636732    chr17-2 1   +   85.6
chr2    231281278   231281455   chr2-2  1   +   83.3

but still 1 questions: "#" mean any pattern I can input, is that right? I didnot use it

• 0 views • link

written 8.8 years ago by Ming Lu •

but still 1 questions: "#" mean any pattern I can input, is that right? I didnot use it

yes, within double quote, you can use any pattern. in this case, line with comment in peak file i.e "#" is not required, so to filter it, "grep -v "#" has been used.

• 0 views • link

written 8.8 years ago by Prakash

why we have to clear lines with #, which didnot impact the intersect manipulation and result? even in homer's pos2bed.pl .txt >.bed, the new .bed file keeps the lines with #

• 0 views • link

written 8.8 years ago by Ming Lu •

you can further shorten the code:

cut -f 1-6 peaks.txt | awk '{print $2,$3,$4,$1,$5}' OFS="\t"

cut will take range and awk can take delimiter to all columns. IMO, that much code is not necessary. Please try the following:

OP:

cut -f 1,2,3,4,6 peaks.txt | awk '{print $2"\t"$3"\t"$4"\t"$1"\t"$5}' >peak1.bed

New code if you have lines with #:

grep -v "#" peak.txt |cut -f 2-4,1,6 > peak1.bed

New code if you do not have lines with #:

cut -f 2-4,1,6 peak.txt > peak1.bed

• 0 views • link

written 8.8 years ago by cpad0112

grep -v "#" peak.txt |cut -f 2-4,1,6 > peak1.bed

Actually, using this code, order of column will not be changed. So, yes below shorter code which you mentioned will solve the purpose.

cut -f 1-6 peaks.txt | awk '{print $2,$3,$4,$1,$5}' OFS="\t"

• 0 views • link

written 8.8 years ago by Prakash

oops...I didn't see 5th column missing.

$ cut -f 1-6 peaks.txt | awk '{print $2,$3,$4,$1,$5}' OFS="\t"

should be

$ cut -f 1-4, 6 peaks.txt | awk '{print $2,$3,$4,$1,$5}' OFS="\t"

• 0 views • link

written 8.8 years ago by cpad0112

Read the documentation!!

-wa option to report the original A feature

One can force bedtools intersect to report the original “A” feature when an overlap is found. As shown below, the entire “A” feature is reported, not just the portion that overlaps with the “B” feature.

• 0 views • link

written 8.8 years ago by BioinfGuru

Thank you , I know it, however pos2bed.pl will cut the column 6 from txt file to bed file. Even when I transform bed to txt again, there is still no column 6.

• 0 views • link

written 8.8 years ago by Ming Lu •