ATAC-seq Analysis Pipeline

Overview

This pipeline processes ATAC-seq data, including trimming, mapping, peak calling, and quality control. It integrates CHIP-Atlas data and generates final outputs like peaks and normalized BigWig files.

---

Requirements

Software:

• cutadapt - For adapter trimming.
• bwa - For mapping reads to the reference genome.
• samtools - For BAM processing.
• macs2 - For peak calling.
• aria2c - For threaded downloads from CHIP-Atlas.
• deeptools - For calculating normalized BigWig files and FRiP scores.
• fastqc - For FASTQ quality control.
• multiqc - For aggregating QC reports.

Python Libraries:

• pandas
• pysam
• deeptools

---

Folder Structure

• config/: Configuration files (e.g., config.yaml).
• data/: Raw FASTQ files or SRR files for download.
• trimmed/: Trimmed FASTQ files after adapter removal.
• results/: Processed outputs including mapping, peaks, and bigwig files.
• qc/: Quality control outputs (e.g., FASTQC, flagstat, stats, and FRiP).

---

Input File

The pipeline requires a sample metadata file in TSV format. Below is an example:

Name	Unit	Library	Fastq1	Fastq2	GSM
LNCaP_ATAC	r1	Paired	/groups/lackgrp/raw_data/LNCaP/ATAC-seq/ATAC_16h_R1_1.fq.gz	/groups/lackgrp/raw_data/LNCaP/ATAC-seq/ATAC_16h_R1_2.fq.gz	-
LNCaP_ATAC	r2	Paired	/groups/lackgrp/raw_data/LNCaP/ATAC-seq/ATAC_16h_R2_1.fq.gz	/groups/lackgrp/raw_data/LNCaP/ATAC-seq/ATAC_16h_R2_2.fq.gz	-

• Name: Sample name.
• Unit: Unit or replicate identifier.
• Library: Single or paired-end sequencing.
• Fastq1: Path to raw FASTQ file (R1).
• Fastq2: Path to raw FASTQ file (R2 for paired-end data, - for single-end).
• GSM: GEO sample ID (optional, used for CHIP-Atlas integration).

---

Configurations

The configuration file (config/config.yaml) specifies pipeline parameters. Below is an example:

SAMPLES: config/samples.tsv

OUTPUT:
  REF: hg38
  RUN:
    QC: True
    PEAKS: True
    BWS: True
    CHIPATLASBED: False
    CHIPATLASBIGWIG: False
  BW_NORMALIZATIONS:
    - rawcount
    - FPM
  BAMPROCESS_PARAMS: -q 30
  MACS_THRESHOLD: 0.01
  IDR_THRESHOLD: 0.05
  CHIPATLASBED_THRESHOLD: '05'

REFERENCES:
  hg38:
    FA: /groups/lackgrp/genomeAnnotations/hg38/hg38.fa
    BWA_IDX: /groups/lackgrp/genomeAnnotations/hg38/hg38.bwa.idx
    CHROM_SIZES: /groups/lackgrp/genomeAnnotations/hg38/hg38.chrom.sizes
    BLACKLIST: /groups/lackgrp/genomeAnnotations/hg38/hg38-blacklist.v2.bed

---

Running the Pipeline

1. Configure the Pipeline

Edit the config/config.yaml file to specify:

• Reference genome details.
• Output directory structure.
• Flags to enable or disable specific steps (e.g., QC, peak calling, BigWig normalization).

Ensure that the config/samples.tsv file is properly formatted with the sample information.

---

2. Slurm Profile

The pipeline uses a Slurm cluster via the profile/ directory. The config.yaml for the Slurm profile should include the following:

cluster:
  mkdir -p logs/{rule} &&
  sbatch
    --partition={resources.partition}
    --cpus-per-task={threads}
    --mem={resources.mem_mb}
    --job-name=smk-{rule}-{wildcards}
    --output=logs/{rule}/{rule}-{wildcards}-%j.out
default-resources:
  - partition=normal,big-mem,long,express
  - mem_mb=700000
  - disk_mb=1024000
restart-times: 1
max-jobs-per-second: 10
max-status-checks-per-second: 1
local-cores: 1
latency-wait: 60
jobs: 12
keep-going: True
rerun-incomplete: True
printshellcmds: True
scheduler: greedy
use-conda: True

• Cluster Resources: Adjust memory (mem_mb), disk (disk_mb), and partition names according to your Slurm setup.
• Logging: Logs for each rule are stored in logs/{rule}/.

---

3. Submission Script

Use the following run_pipeline.sh script to submit the pipeline to the Slurm cluster. The script activates the required conda environment and runs Snakemake with the specified profile.

#!/bin/bash
#SBATCH -c 64
#SBATCH --mem 720GB
#SBATCH -p long,big-mem,normal,express

source ~/.bashrc
conda activate atacseq

snakemake --profile profile/

---

4. Submit the Pipeline

Run the following command to execute the pipeline:

sbatch run_pipeline.sh

This will:

• Automatically submit jobs to the Slurm cluster.
• Use the configuration specified in the profile/config.yaml file.
• Execute all defined rules in the pipeline.