Can I successfully go from a bed file (narrowpeak) back to a bam file containing the mapped ATAC-seq reads?

Question

[Jillian]

Hello, I am very new to the atac-seq analysis pipeline. I am going to try footprinting analysis of some published ATAC-seq data using TOBIAS. I need the BAM file containing mapped reads, though I only have a few files available to me. Can I go from a bed file (narrow peak) back to the bam file I mentioned? (Off hand, it doesnt seem to be possible since the narrow peak only contains peak metadata)

Alternatively, can I go from a normalized bigwig signal file back to the bam file I need (i.e. mapped read bam file)? The bigwig file is normalized coverage of 1-100 bp ATAC-seq fragments

Here is a line from bed file:

2L 5135 5369 . 105 . rtracklayer sequence_feature . name=V11_3_INTACT_peak_1;signalValue=2.16438;pValue=11.82981;qValue=10.59647;peak=132

Thanks for your patience and help!

[Pierre Lindenbaum]

I'm sorry but it feels like asking "I've got a steak, can I get the cow ?"

[Alex Reynolds]

Close, but no CIGAR. You'll need more data.

[Jillian]

thank you for these (somewhat polite and helpful) comments---as said in my question, that is what I figured. I have asked the lab for their upstream data (FASTQ, bam)

Answer 1

[seidel]

I need the BAM file containing mapped reads, though I only have a few files available to me.

Among the few files you have, if they are explicitly just the BED and bigWig files, then these are not even the steak. They're more like the dimensions (BED) of the shadow (bigWig) of the cow. You can't get the cow from the shadow (as has been pointed out). If the data is published, can't you get access to the fastq files from a repository? With those, you can regenerate BAMs or whatever else you need.

[Jillian]

thanks-although the needed files are not in a database, I contacted the lab and they said they would send the fastq files.

Answer 2

[ATpoint]

Plainly and simply: No. The mapping information is completely lost in these files.