How to perform analysis for repetitive elements in ATAC-Seq data?

Question

[Deepak Tanwar]

Hello!

I have questions about performing repeat-elements (RE) analysis in ATAC-Seq data.

So, I have RE information from repeatmasker

RE elements table screenshot

Now, I have bam files from ATAC-Seq.

I think I have 2 approaches to know which repeat-elements have less or more chromatin accessibility (depending on how I want to look at the RE):

1. Read the counts in the windows of each RE transcript and perform differential analysis.
2. Read the counts at the sub-family level and perform differential analysis.

Question:

Is my approach correct, statistically?

[ATpoint]

Out of interest, can you map reads uniquely to these repeats? I imagine you have a lot of multimappers and therefore only few uniquely / reliable alignments to them, and by this a sparse matrix of counts?

[Deepak Tanwar]

I do not have a sparse matrix for ATAC-Seq data. After filtering out the RE with low counts, I still left with approx. 1.5M RE. I did not filter out reads with MAPQ less than 30.

But, for RNA-Seq, one has to allow for multi-mapping.

For point number 2 in my question: https://doi.org/10.1186/s13100-019-0192-1