nucMACC

Original paper - https://doi.org/10.1101/2022.12.29.521985

Introduction

nucMACC is an automated analysis pipeline for the analysis of nucleosome positions, accessibility and stability. The pipeline contains two main workflows:

1. MNaseQC for QC and exploratory analysis
2. nucMACC for analysis of nucleosome positions, accessibility and stability

Given trimmed paired-end sequencing reads in fastq format, this pipeline will run:

• MNaseQC and nucMACC

  1. QC using FastQC on fastq files
  2. Alignment using Bowtie2 on fastq files
  3. QC using Qualimap on aligned fragments
  4. Fragment size distribution plot
  5. Group the fragments by size using 'deepTools alignmentSieve' and optionally filter blacklisted regions
     1. Mono-nucleosome (140 - 200 bp)
     2. Sub-nucleosome (< 140 bp)
  6. Report fragment statistics of each processing step
  7. Create nucleosome maps of Mono- and Sub-nucleosomes using DANPOS
  8. Optionally create TSS profiles using deepTools computeMatrix
  9. Summary reports using MultiQC

• MNaseQC specific

  1. PCA of nucleosome maps using deepTools
  2. Correlation analysis using deepTools

• nucMACC specific

  1. Pool all mono-nucleosome samples
  2. Obtain mono-nucleosome positions from pooled samples and sub-nucleosome positions from lowest MNase concentration using DANPOS
  3. Get GC content of nucleosome positions using bedtools genomecov
  4. Count number of fragments using featureCounts
  5. Remove nucleosome positions with low fragment count (mono-nucleosmes < 30 and sub-nucleosomes < 5)
  6. Calculate (sub-)nucMACC score using linear regression analysis
  7. Correct for MNase GC-bias using LOWESS
  8. Identify hyper-/hypo-accessible nucleosomes or unstable and non-canoncical nucleosomes.

nucMACC is meant to run on pooled replicates in fastq format, whereas MNaseQC uses single replicates. As the MNaseQC and the nucMACC workflow have several steps in common, it is recommended to run first MNaseQC and report the fragment size selected bam files using --publishBamFlt. Then setting --bamEntry option, a shorter version of the nucMACC workflow can be run using the generated bam files as input. Here in an additional step at the beginning replicates are pooled.

Get started

Requirements

• Docker and nextflow are required to run the nucMACC pipeline. Additional software used in the pipeline is packaged in Docker container and will be automatically downloaded during the first execution of the pipeline.
• The pipeline is compatible with all computational infrastructures. Executing the pipeline on cloud or HPC systems may require to adapt the nextflow.config.
• For large reference genomes the pipeline requires at least 32 GB memory and default settings allocate 45 GB memory to demanding processes. Memory usage can be adjusted using the option --high_memory or directly in the nextflow.config.
• The pipeline requires paired-end sequencing data
• Bowtie2 index of reference genome, reference genome in fasta format and the length of the mappable genome.
• For details about the GTF file format required for TSS plots, see the documentation of deeptools computeMatrix. For further customisation of TSS plots we recommend direct use of the deepTools package and the bigwig files provided by the pipeline.

Installation

You can obtain the pipeline directly from GitHub:

git clone https://github.com/uschwartz/nucMACC.git

Test pipeline

The pipeline comes with a ready-to-use test data set.

nextflow run path2nucMACC/nucMACC --test

Usage

We recommend to use first the MNaseQC workflow and specifying --publishBamFlt. Then take the output and run nucMACC with --bamEntry option.

To execute the pipeline a samplesheet is required. The content depends on the workflow to execute. See examples in the toyData folder.

Workflow:

• MNaseQC (example toyData/input_replicates.csv)

Sample_Name,path_fwdReads,path_revReads,MNase_U
H4_rep1_6.25U_cut,/toyData/H4_rep1_6.25U/H4_rep1_6.25U_cut_1.fastq.gz,/toyData/H4_rep1_6.25U/H4_rep1_6.25U_cut_2.fastq.gz,6.25
H4_rep2_6.25U_cut,/toyData/H4_rep2_6.25U/H4_rep2_6.25U_cut_1.fastq.gz,/toyData/H4_rep2_6.25U/H4_rep2_6.25U_cut_2.fastq.gz,6.25
H4_rep1_100U_cut,/toyData/H4_rep1_100U/H4_rep1_100U_cut_1.fastq.gz,/toyData/H4_rep1_100U/H4_rep1_100U_cut_2.fastq.gz,100
H4_rep2_100U_cut,/toyData/H4_rep2_100U/H4_rep2_100U_cut_1.fastq.gz,/toyData/H4_rep2_100U/H4_rep2_100U_cut_2.fastq.gz,100

Each row represents a pair of fastq files. Here unique sample names are required.

• nucMACC --bamEntry (example toyData/sub_input.csv)

Sample_Name,replicate,path_mono,path_sub,MNase_U
H4_6.25U,rep1,/toyData/monoNuc/H4_rep1_6.25U_cut_mono.bam,/toyData/subNuc/H4_rep1_6.25U_cut_sub.bam,6.25
H4_6.25U,rep2,/toyData/monoNuc/H4_rep2_6.25U_cut_mono.bam,/toyData/subNuc/H4_rep2_6.25U_cut_sub.bam,6.25
H4_100U,rep1,/toyData/monoNuc/H4_rep1_100U_cut_mono.bam,/toyData/subNuc/H4_rep1_100U_cut_sub.bam,100
H4_100U,rep2,/toyData/monoNuc/H4_rep2_100U_cut_mono.bam,/toyData/subNuc/H4_rep2_100U_cut_sub.bam,100

Each row represents a pair of fastq files. Rows with the same sample name are considered technical replicates and pooled automatically. Only numerical values are allowed in the last column MNase_U. Duration of MNase experiment could be used as well, if the MNase concentration was constant in the experiments, but the time of digestion differed. It is recommended to use the output of MNaseQC workflow, which can be obtained specifying --publishBamFlt. However, it is as well possible to enter the pipeline at this point with manually processed bam files.

• nucMACC (example toyData/input.csv)

Sample_Name,path_fwdReads,path_revReads,MNase_U
H4_rep1_6.25U_cut,/toyData/H4_rep1_6.25U/H4_rep1_6.25U_cut_1.fastq.gz,/toyData/H4_rep1_6.25U/H4_rep1_6.25U_cut_2.fastq.gz,6.25
H4_rep1_100U_cut,/toyData/H4_rep1_100U/H4_rep1_100U_cut_1.fastq.gz,/toyData/H4_rep1_100U/H4_rep1_100U_cut_2.fastq.gz,100

Each row represents a pair of fastq files. In case of several replicates per MNase titration point, the fastq files need to be pooled before starting the pipeline. Only numerical values are allowed in the last column MNase_U. Duration of MNase experiment could be used as well, if the MNase concentration was constant in the experiments, but the time of digestion differed.

Execute:

• MNaseQC

nextflow run path2nucMACC/nucMACC \
        --analysis 'MNaseQC' \
        --csvInput 'sample_sheet.csv' \
        --outDir <OUTDIR> \
        --genomeIdx 'Bowtie2Index/genome' \
        --genomeSize 119481543 \
        --genome 'genome.fa' \
        --publishBamFlt \
        --blacklist 'blacklisted_regions.bed'  \
        --TSS 'genes.gtf'

All options, except --publishBamFlt,--blacklist, and --TSS, are required.

• nucMACC with --bamEntry

nextflow run path2nucMACC/nucMACC \
        --analysis 'nucMACC' \
        --csvInput 'sample_sheet.csv' \
        --outDir <OUTDIR> \
        --genomeIdx 'Bowtie2Index/genome' \
        --genomeSize 119481543 \
        --genome 'genome.fa' \
        --bamEntry \
        --TSS 'genes.gtf'

All options, except --TSS, are required.

Get help

nextflow run path2nucMACC/nucMACC --help

Contact

Please log all issues/suggestions on the nucMACC GitHub page: https://github.com/uschwartz/nucMACC/issues

Uwe Schwartz: uwe.schwartz@ur.de

Cite

Sara Wernig-Zorc et al. ,nucMACC: An MNase-seq pipeline to identify structurally altered nucleosomes in the genome.Sci. Adv.10,eadm9740(2024).DOI: 10.1126/sciadv.adm9740 (https://doi.org/10.1126/sciadv.adm9740)