ATAC-seq

Updated by Hongjiang on 09/17/2022

Assay for Transposase-Accessible Chromatin using sequencing(ATAC-seq) is a sequencing technology to detect genome-wide chromatin accessibility.

The genomic DNA of eukaryotic cells is entangled in histones, which is inaccessible in this case, while actions such as transcription make this DNA accessible, i.e., it can interact with cis-transons. ATAC-seq, on the other hand, usually uses Tn5, a transposase enzyme, to process the genome, which cut the accessible regions randomly into fragments of varying length (length determined by processing time, normally distributed). Then these fragments are built and sequenced to obtain ATAC-seq sequencing information.

Upstream Analysis

ENCODE-DCC/atac-seq-pipeline

Install Conda

Refer to Tools/Conda

Install ATAC-seq pipeline(v1.10.0)

This tutorial guides the installation of v1.10.0. v2 and later versions have different installation methods due to caper version changes, which does not apply this tutorial.

The upstream analysis results for v1 and v2 are literally the same, except for the dependency package upgrade.

Pipeline Installation

1.Download from Github:

wget https://github.com/ENCODE-DCC/atac-seq-pipeline/archive/refs/tags/v1.10.0.zip
unzip v1.10.0.zip

2.Uninstall previous version of pipeline's conda environment (if required):

bash scripts/uninstall_conda_env.sh

3.Install the conda environment for v1.10.0:

🕒Takes about 2 hours.

bash scripts/install_conda_env.sh

4.Download gene reference file genome.tsv：

5.

JDK(java) Installation

1.Activate conda env:

conda activate encode-atac-seq-pipeline

2.Install JDK:

conda install -c conda-forge openjdk=8

3.Check JDK version:

java -version

The output should include Openjdk version "1.8.*".

Caper Installation

ENCODE-DCC/caper

1.Activate conda env:

conda activate encode-atac-seq-pipeline

2.Install caper:

pip install caper==1.6.3

⚠️Note: You need to specify the caper version as 1.6.3, no version >2 can be installed.

3.Initialize caper:

If running on a local server (local computer without cluster engine), run:

caper init local

If running on a cluster, perform SGE (Sun GridEngine cluster) initialization:

caper init sge

The caper initialization also supports slurm, pbs, lsf, etc.

This will initialize the runtime environment in ~/.caper and download the two jar packages Cromwell and Womtool to the ~/.caper/cromwell_jar/ and ~/.caper/womtool_jar/ folders, and generate the ~/.caper/default. conf configuration file.

Different versions of Cromwell and Womtool can be specified by changing the configuration file.

All versions of Cromwell/Womtool

Downgrade tbb

Higher versions of tbb doesn't work with samtools:

conda install tbb=2020.2