ChIP-Seq density plot of the TSS regions?

Question

[Susmita Mandal]

Hello everyone,

I have done ChIP-Seq analysis alllele specific way somehow, I think. I aligned the ChIP-Seq data to both the parental genomes and then I merged the aligned BAM files and I did peak calling on the merged BAM file using MACS2. I got a BED file and a bedgraph file from it. I intersected that bedgraph file with the individual parental VCF files and got allele specific bedgraph files that converted to bigwig and I visualized in the IGV. Now I want to quantify those data and plot density plots of the TSS of the genes allele specific wise. After reading many papers and blogs and biostars, I got few words like spanning, windows, variable step and deeptools and what not. I know how to use deeptools but I don't think its giving results as I want. So can anyone make me explain step by step as how to do it. Any help is appreciated.

Thanks in advance.

Susmita

[Friederike]

I know how to use deeptools but I don't think its giving results as I want

How is that? I.e. what type of result to you get and what do you not like about it?

Answer 1

[i.sudbery]

The tool bam2geneprofile from cgat can produce TSS metagene plots from bigwig files.

I don't really understand how you converted a MACS2 bigwig into allele specific bigwigs, but if thats really what you've got, then after installing cgat simply run:

cgat bam2geneprofile --method=tssprofile --bigwigfile=maternal_signal.bw --gtf-file=gene-models.gtf.gz --output-pattern=maternal.%s
cgat bam2geneprofile --method=tssprofile --bigwigfile=paternal_signal.bw --gtf-file=gene-models.gtf.gz --output-pattern=paternal.%s

See cgat bam2geneprofile --help for ways in which profiles can be normalised.

I've also found that removing overlapping genes can make a big difference to metagene profiles. I don't know if anyone has a better way to do this, but I do it using a combination of cgat and bedtools with:

cgat gtf2gtf --method=merge-transcripts -I gene-models.gtf.gz \
                 | cgat gff2bed --is-gtf -L gene-models.filtered.log \
                 | bedtools slop -l 1250 -r 1250 \
                                 -s -i - -g contigs.tsv \
                 | sort -k1,1 -k2,2n \
                 | bedtools merge -c 4 -o count \
                 | awk '$4>1' \
                 | bedtools intersect -v -a gene-models.gtf.gz -b - \
                 | bgzip > gene-models.filtered.gtf.gz