atpoint / atac_chip_preprocess Goto Github PK

samplesheet input and validation

allow multiple libraries per sample and auto-merge

- joint peak calling over all BAM files

• add NGS blacklists for mm10 and hg38 as assets

• add experimental groups to samplesheet
• get consensus peaks per group
• in samplesheet, allow a field "exclude" that if set the sample is not considered, e.g. when QC turns out to be bad one simply alters the samplesheet and uses -resume so the consensus peaks etc would be based on the good samples only
• make normalized bigwig files, also averaged by group

add:
-bigwig smoothed by 100bp
-cutsite bigwigs

both unnormalized so can be easily scaled after inspecting the QC to decide which samples are ok and will be included into downstream

Hello @ATpoint , I hope this is a good place for me to ask a question regarding your pipeline.

In some ATAC-seq pipelines, fragments are size filtered (< 100 bp) to select for NFR, then fed into MACS2 for peak calling while some pipelines feed all the fragments.

I was wondering what your rationale was for not selecting for NFR (I tried to go through your source code, and please correct me if I am wrong).

Thanks! (and your help on biostars is always very helpful, thanks so much)

Write samples with zero peaks as FRiP 0 into frip output.