Saved input bams for MACS2?

[bhanco]

Hi,

I am looking for the saved input bams for MACS2, these should have undergone all preprocessing stems including spike-in normalization (by downsampling reads to IGG or by spike-in reads to 10,000 I assume).

The documentations seems to suggest that these are saved.
From: https://nf-co.re/cutandrun/3.2.2/docs/output/
"The pipeline will output the .bam files with index and samtools stats for only the final set by default. For example, the full pipeline will only output picard duplicates processed files as this is the final step before peak calling."

But I can't find these final input bam files
02_alignment/bowtie2 only has deduped and marked duplicate files for the IGG. (dedup not expected to run on sample)

And from there 03_peak_calling/01_bam_to_bedgraph starts with .bedGraphs.

I ran with --saved_aligned_intermed true

But still can't find the files.

In an earlier comment an author response suggested MACS2 takes .bedGraph files. However, the input to macs2 seems to want a bam based on variable names.
"input:
tuple val(meta), path(ipbam), path(controlbam)
val macs2_gsize"

And based off the call in the main workflow:
MACS2_CALLPEAK_IGG (
ch_bam_paired,
params.macs_gsize
)
Where as seacr takes .bedGraph
SEACR_CALLPEAK_IGG (
ch_bedgraph_paired,
params.seacr_peak_threshold
)

The input for seacr I assume are the .bedGraph files in the 03_peak_calling dir I assume. But where are the inputs for MACS2?

[johnvusich]

Hi @bhanco,

You are right that MACS2 uses BAM as input. Spike-in normalization occurs by computing a scaling factor during the bam_to_bedgraph step of the pipeline. Therefore, the inputs for MACS2 are the BAM files.

Using the nf-core/cutandrun test data, here is the script used for MACS2 peak calling:

macs2 \ callpeak \ --nomodel --shift -75 --extsize 150 --keep-dup all -q 0.01 \ --gsize 2700000000.0 \ --format BAMPE \ --name h3k27me3_R1.macs2 \ --treatment h3k27me3_R1.target.linear_dedup.sorted.bam \ --control igg_ctrl_R1.target.linear_dedup.sorted.bam

The input BAM files should be located in results/02_alignment/bowtie2/target/linear_dedup.