Is strand important when it comes to DNAse and ChIP-Seq peak calling?

Question

[ashah90]

Hello everyone,

I have recently been using data from the Epigenomics Roadmap Project and am a little confused. I have been downloading the data from http://egg2.wustl.edu/roadmap/web_portal/processed_data.html#ChipSeq_DNaseSeq under section "c. Peak Calling" for NarrowPeak data. The data will have the following format:

chrom    chromStart    chromEnd    rank    score    strand    signal_value    pvalue    qvalue    peak
chr20    30298908    30300550    Rank_1    423    .    11.54602    42.39392    34.03034    1197

As you can see the "strand" column is left blank. What I am trying to do is check whether peaks in different cell lines and histone marks map to regulatory regions within the genome. I often will use the 2000 bp and 500 bp downstream criteria when defining a promoter region, but taking strand into account and with the way UCSC defines their txstart and txend fields, for negative strand genes, I have to go 2000 bp downstream the TSS rather than upstream. This brings up the question as to whether these peaks are being mapped on the positive or the negative strand, or does it not matter? Are the coordinates such that it is the positive strand orientation?

Thanks you for your help

Answer 1

[Ying W]

The concept of strand on a gene comes from transcription. You could have a gene transcribed from the positive strand or negative strand of DNA. You can tell from the mRNA which strand the gene was transcribed from. For DNAse-seq and ChIP-seq you are assaying a position in the genome (thus no strand). This assay is determines where the DNA is less compact or where a protein is binding to DNA (no concept of +/- strand)

Answer 2

[Ming Tommy Tang]

There is no strandness for ChIP-seq and Dnase-seq peaks. When counting reads in the peaks, both + reads and - reads will be counted.

[Bioinformatist Newbie]

If DNAse-seq peaks are called by MACS, is there any standard threshold for annotating genes as accessible or not accessible (based on the peak tag density score)?

[Ming Tommy Tang]

There is no standard threshold. You need to set it as long as it makes biological sense.

[Bioinformatist Newbie]

Can you point me out towards a publication mentioning the reason behind the threshold they selected?