How to make an average conservation plot from ChIP-seq data

[crazyhottommy]

Hi all,

I want to know how to make an average conservation plot
http://ceas.cbi.pku.edu.cn/ this website can make something I want, but it can only plot
a set of regions at one time, I want to plot two sets of regions, and compare them.

" 2. GC content and evolutionary conservation of each ChIP-region and their
average. CEAS uses PhastCons conservation scores from UCSC Genome
Bioinformatics, which is based on multiz alignment of human, chimp, mouse,
rat, dog, chicken, fugu, and zebrafish genomic DNA. CEAS generates thumbnail
conservation plot for each ChIP-region and the average conservation plot for
all the ChIP-regions, which can be directly used in ChIP-chip biologists'
manuscript.
"


Any python scripts or bioconductor package can do it?

Thanks

[paolo.kunder]

Hi,

I encountered the same problem few weeks ago,

It seems that there are not many "expert" regarding PhastCons in this forum.

I explain you what I did,

I took my Chip-seq regions, in bed file format, intersect them with PhastCons element in (mouse example)

http://hgdownload-test.cse.ucsc.edu/goldenPath/mm9/phastCons30way/

and calculated for each genomic position the score of conservation.
Then averaged all the score.

Consider that you will need to transform a bit the PhastCons file format,

for example:

original format ( just replaced some words with empty)

chrom=chr1 start=3000306
0.006
0.010
0.014
chrom=chrX start=40000306
0.014
chrom=chr9 start=80000306 
0.1
0.2

processed format

chr1 3000306 3000307 0.006
chr1 3000307 3000308 0.010
chr1 3000308 3000309 0.019
chrX 40000306 40000307 0.014
chr9 80000306 80000307 0.1
chr9 80000307 80000308 0.2

with the following script:

awk '/^chrom/{split($1,a,"=");split($2,b,"=");next} { printf "%s\t%10d\t%10d\t%f\n",a[2],b[2],b[2]+1,$1;b[2]++}' filename

Another possibility that I am investigating is using circos

http://circos.ca/

but You will need to study first how the software works,

Cheers,
Paolo

[crazyhottommy]

Thank you so much!!

Hi,

I encountered the same problem few weeks ago,

It seems that there are not many "expert" regarding PhastCons in this forum.

I explain you what I did,

I took my Chip-seq regions, in bed file format, intersect them with PhastCons element in (mouse example)

http://hgdownload-test.cse.ucsc.edu/goldenPath/mm9/phastCons30way/

and calculated for each genomic position the score of conservation.
Then averaged all the score.

Consider that you will need to transform a bit the PhastCons file format,

for example:

original format ( just replaced some words with empty)

chrom=chr1 start=3000306
0.006
0.010
0.014
chrom=chrX start=40000306
0.014
chrom=chr9 start=80000306 
0.1
0.2

processed format

chr1 3000306 3000307 0.006
chr1 3000307 3000308 0.010
chr1 3000308 3000309 0.019
chrX 40000306 40000307 0.014
chr9 80000306 80000307 0.1
chr9 80000307 80000308 0.2

with the following script:

awk '/^chrom/{split($1,a,"=");split($2,b,"=");next} { printf "%s\t%10d\t%10d\t%f\n",a[2],b[2],b[2]+1,$1;b[2]++}' filename

Another possibility that I am investigating is using circos

http://circos.ca/

but You will need to study first how the software works,

Cheers,
Paolo

[szaman]

Making the step size 200 bp

Hello Paolo,

Your code is extraordinarily useful, however, I am trying to make the step size 200 bp and while doing so, I want to average the phastCon score over each 200 bp region. I am not even sure how to get started (really not even sure which language to use, I only know rudimentary unix, perl, & R). As much as I love one-liner codes, I do not think that it possible in this situation. I am working with hg38.phastCons20way.wigFix (phastcon scores calculates for human genome wide based on 19 species alignments) and it is a huge file. Any guidance would be much appreciated.