MNase-seq Data Analysis and Processing

Question

[cbio]

I was hoping someone could point me in the right direction for MNase-seq analysis. I just recently got my hands on a dataset, and after a great deal of looking around, I'm having trouble finding a proper protocol.

Could I get a general outline? I know some sequencing types don't like having their fastq files filtered because it messes with downstream analysis and that sort of thing.

So something like:

1) Filter fastq

2) Map to genome using bowtie2

3) Filter sam/bam for duplicates, trim reads

4) call peaks with macs2

and so on. I am attempting to identify nucleosome positions in mm9 genome but would like to read a few papers that go over the most commonly used tools for mapping and peak calling and so forth.

Answer 1

[dariober]

I have limited experience with MNAse-Seq data and it seems to me it's not the cleanest data ever seen... Anyway, the first steps are common to most *-Seq protocols:

• Trim reads to remove adapters (e.g. with cutadapt)
• Align with your favourite tool, I use bwa mem but it should make little difference.
• Remove duplicate reads and reads mapped ambiguously (e.g. with samtools view -F 3844 -q 5 ...)

So far so good. Now problems begin, at least my experience. MNAse-Seq doesn't really produce "peaks" at most you see bumps a bit everywhere and MACS doesn't do a good jobs. Recently I used iNPS and looks ok-ish, with a bit of imagination. As a sort of QC you could look at the average profile around CTCF sites as you should see a valley centred on CTCF and regular "waves" on its sides (just search google or pubmed for CTCF+nucleosome). What do to next depends on your exact question...

Answer 2

[Carlo Yague]

I suggest you take a look at the DANPOS tool and paper*. It is especially designed for MNase-seq analysis with interesting features such as clonal reads removal, fragment size estimation/correction and read length adjustment.

The tool can take as input aligned bam (better not remove duplicates, DANPOS has its own filter) and output coverage files and peak position files. It also defines "dynamic nucleosomes" if you compare two conditions. Those nucleosomes show differences of either occupancy, fuzziness or position accross conditions. The three classes of dynamic changes are separately called and may reflects different biological realities.

Kaifu Chen, etc. DANPOS: Dynamic Analysis of Nucleosome Position and Occupancy by Sequencing. Genome Research. 2012.*

PS : DANPOS is now part of a bigger workflow called DANPOS2 as the DPOS function.