Skip to content

Question regarding region-based DMR analysis #718

Description

@tinatucci

Hi,

Thank you for developing a fantastic tool. I had a question regarding DMR analysis with supplied target regions. We have performed a sliding-window like analysis on our datasets which worked, but we need to know the number of unmod/modified sites within these small regions contributing to the call of 'differentially modified region'. We are, for instance, more interested in regions containing at least 3 sites versus those with a single site. Our end goal is to find small (~200 bp) regions with 3 or greater sites that are differentially modified between two conditions. We have done this using replicates (3 for each condition).

This is the current approach:

    modkit dmr pair \
        --ref ${ref} \
        -r ${regions_bed} \
        ${aArgs} \
        ${bArgs} \
        --base ${mod.base} \
        --single-code ${mod.code} \
        --threads ${task.cpus} \
	--header \
        --log-filepath ${comp_name}_${mod.name}.log \
        -o ${comp_name}_${mod.name}_dmr.bed

The resulting header is:
chrom start end name score strand a_counts a_total b_counts b_total a_mod_percentages b_mod_percentages a_pct_modified b_pct_modified effect_size cohen_h cohen_h_low cohen_h_high

Is there a way to know the number of sites within DMR's of interest? Or perhaps to force the analysis to report only those with >3 sites? Advice on the best way forward to achieve this would be great.

Many thanks!

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions