UF Health Cancer Center Bioinformatics Hub
Analysis of Omics Data by Assay
ATAC-seq

ATAC-seq

ATAC-seq Analysis Overview

ATAC-seq analysis aims to identify regions of open chromatin and detect differential accessibility between conditions. The workflow involves: (1) preprocessing and peak calling, (2) differential accessibility analysis, and (3) biological interpretation through annotation and motif analysis.

The analysis path depends on your research questions and data characteristics, with key decision points around peak calling strategies and differential analysis methods.

Analysis Workflows

Primary Processing: nf-core/atacseq Pipeline

What it does: The nf-core/atacseq pipeline handles complete preprocessing from raw reads through peak calling, including adapter trimming, alignment (bowtie2), duplicate removal, QC reporting, and peak calling with MACS2.

Input: Raw FASTQ files
Output: Aligned BAM files, QC reports, peak calls (BED files), fragment size distributions

Recommendation: Start with nf-core/atacseq for all projects - you can always use alternative peak callers or analysis methods on the BAM file outputs.

Peak Calling Strategy

Use Default MACS2 Peaks from nf-core/atacseq

Best for: Most ATAC-seq projects with standard experimental designs

Proceed to: Differential analysis with peak-based methods

Alternative Peak Callers

When to consider: If MACS2 peaks don’t capture your regions of interest

Options:

  • Genrich: Optimized for ATAC-seq data, better replicate handling
  • HMMRATAC: Uses fragment size information for improved peak detection

Input: BAM files from nf-core/atacseq
Output: Alternative peak calls (BED files)

Skip Peak Calling Entirely (CSAW Approach)

When to consider: Unbiased genome-wide analysis, broad accessibility changes, or when peak calling may introduce bias

Method: CSAW uses sliding windows across the genome for differential analysis

Input: BAM files from nf-core/atacseq
Output: Window-based differential accessibility results

Differential Accessibility Analysis

Peak-Based Analysis

Recommended: atacreportR - UF-developed application for automated differential analysis and reporting

Alternative: DiffBind for R-based analysis with custom parameters

Input: Peak files (BED) + BAM files + sample metadata
Output: Differential accessibility results, statistical summaries, visualizations

Window-Based Analysis (CSAW)

Method: Direct differential testing on genomic windows using edgeR

Input: BAM files + experimental design
Output: Window-level differential accessibility with specialized FDR control

Biological Interpretation

Tools: ChIPseeker for annotation, motif analysis tools

Choosing Your Analysis Strategy

Peak-Based vs. Window-Based Analysis

Peak-Based Advantages:

  • Faster computation, standard approach
  • Results interpretable as discrete regulatory elements
  • Integrated reporting available (atacreportR)
  • Extensive tool ecosystem

Peak-Based Limitations:

  • Peak calling may miss broad regions or introduce bias
  • Assumes discrete accessible regions

Window-Based (CSAW) Advantages:

  • Unbiased genome-wide analysis
  • Better for broad or diffuse changes
  • Specialized statistical methods for ChIP-seq-like data

Window-Based Limitations:

  • Computationally intensive
  • Complex result interpretation
  • Manual R implementation required

Standard Workflow

Best for: Most ATAC-seq projects

  1. Run nf-core/atacseq pipeline → Complete preprocessing and peak calling
  2. Differential analysis with atacreportR → Automated analysis and comprehensive reporting
  3. Review results → Interactive visualizations and downloadable reports

Custom Peak Calling Workflow

Best for: Projects requiring specialized peak detection

  1. Run nf-core/atacseq pipeline → Get BAM files and QC metrics
  2. Alternative peak calling → Genrich, HMMRATAC, or custom parameters
  3. Differential analysis → DiffBind or atacreportR with custom peaks

Unbiased Discovery Workflow

Best for: Exploratory projects or broad accessibility changes

  1. Run nf-core/atacseq pipeline → Get processed BAM files
  2. CSAW analysis → Window-based differential accessibility
  3. Custom interpretation → Manual R analysis of results

Important Notes

  • Pipeline execution: You run nf-core/atacseq yourself on HiperGator using SLURM
  • atacreportR access: Private UF application requiring network access and coordination
  • Computational resources: Use HiperGator’s SLURM scheduler for all analyses

Getting Started

Submit a support request to discuss your experimental design, choose the appropriate analysis strategy, and obtain access to UF-specific tools like atacreportR.