Skip to content

ASPEN

Table of Contents

1. Outline

ASPEN or Atac Seq PipEliNe is CCBR's pipeline to calls peaks for ATAC-Seq datasets. It currently accepts paired-end Illumina data and calls peak using MACS2 and Genrich peak callers. Below is a brief outline of the steps performed by the pipeline:

  • Trim PE reads with CutAdapt
  • Remove reads aligning to known blacklisted regions, if provided
  • Align reads provided genome using bowtie2. This step generates multiple output files:
  • tagAlign.gz, which is a BED6 format file mainly required for MACS2 peak calling
  • dedup.bam, deduplicated BAM format file which may be required for downstream processing (eg. TOBIAS)
  • qsorted.bam, query sorted BAM file for Genrich peak calling
  • Pre-peakcalling QC metrics:
  • FastQC is run pre- and post-trimming
  • Fragment length distribution is calculated using custom scripts
  • Preseq is run to estimate library complexity
  • Post-peakcalling QC metrics:
  • TSS distributions are calculated for each replicate
  • FRiP (Fraction of Reads in Peaks) is calculated for each replicate
  • FRiPextra calculations are performed if fripextra config files are supplied
    • Fraction of reads in DHS regions
    • Fraction of reads in promoter regions
    • Fraction of reads in enhancer regions
  • Peak calling: Peaks (NarrowPeak format) are called using MACS2 and Genrich. If multiple replicates exist per sample, consensus peaks are called (BED format).
  • Peak annotation: ChIPseeker is used to annotate peaks if the genome is hg38/hg19/mm10
  • Motif enrichment: Motif Enrichment is calculated using HOMER and AME (MEME suite)
  • Report: MultiQC is used to generate a customized final HTML report

2. Runtime details

2.1 Load module on BIOWULF

To clone the repo run:

% module load ccbrpipeliner

This will add aspen to your PATH environmental variable.

NOTE: If not running on BIOWULF, please ensure that the required tools like snakemake, python and singularity are in PATH.

2.2 Create Sample Manifest

Once the data is stored on biowulf, sample manifest TSV (samples.tsv) can be created to have the following columns:

  1. replicateName

  2. sampleName: Multiple replicates can have the same sampleName

  3. path_to_R1_fastq: absolute path is preferred

  4. path_to_R2_fastq: Required! Pipeline currently only supports paired-end data

Note that:

  • symlinks are created for R1 and R2 files from the sample manifest in the results folder. These symlinks have the filenames \<replicateName>.R1.fastq.gz and \<replicateName>.R2.fastq.gz, respectively. Thus, original filenames do not matter and original files do not need to be renamed.
  • replicateName is used as prefix for individual peak calls
  • sampleName is used as prefix for consensus peak calls

NOTE: Optionally, if running differential ATAC please also provide contrasts.tsv in the output folder after running init. This is a simple tab-delimited text file with 2 columns (Group1 and Group2) without any headers.

2.3 Run Pipeline

To get more information about how to run the pipeline simply cd to the above CCBR_ATACseq folder and run

% aspen --help
##########################################################################################

Welcome to
____ ____ ___  ____ _  _
|__| [__  |__] |___ |\ |
|  | ___] |    |___ | \|    v1.0.0

A_TAC_S_eq A_nalysis P_ip_E_li_N_e

##########################################################################################

This pipeline was built by CCBR (https://bioinformatics.ccr.cancer.gov/ccbr)
Please contact Vishal Koparde for comments/questions (vishal.koparde@nih.gov)

##########################################################################################

Here is a list of genome supported by aspen:

  * hg19          [Human]
  * hg38          [Human]
  * mm10          [Mouse]
  * mmul10        [Macaca mulatta(Rhesus monkey) or rheMac10]
  * bosTau9       [Bos taurus(cattle)]

aspen calls peaks using the following tools:

 * MACS2
 * Genrich        [RECOMMENDED FOR USE]

USAGE:
  bash ./aspen -w/--workdir=<WORKDIR> -m/--runmode=<RUNMODE>

Required Arguments:
1.  WORKDIR     : [Type: String]: Absolute or relative path to the output folder with write permissions.

2.  RUNMODE     : [Type: String] Valid options:
    * init      : initialize workdir
    * dryrun    : dry run snakemake to generate DAG
    * run       : run with slurm
    * runlocal  : run without submitting to sbatch
    ADVANCED RUNMODES (use with caution!!)
    * unlock    : unlock WORKDIR if locked by snakemake NEVER UNLOCK WORKDIR WHERE PIPELINE IS CURRENTLY RUNNING!
    * reconfig  : recreate config file in WORKDIR (debugging option) EDITS TO config.yaml WILL BE LOST!
    * reset     : DELETE workdir dir and re-init it (debugging option) EDITS TO ALL FILES IN WORKDIR WILL BE LOST!
    * printbinds: print singularity binds (paths)
    * local     : same as runlocal

Optional Arguments:

--help|-h       : print this help
--genome|-g     : genome eg. hg38
--manifest|-s   : absolute path to samples.tsv. This will be copied to output folder                    (--runmode=init only)
--useenvmod|-e  : use "--use-enmodules" option while running Snakemake. This is for using modules on HPC instead of containers(default).
--singcache|-c  : singularity cache directory. Default is `/data/${USER}/.singularity` if available, or falls back to `${WORKDIR}/.singularity`.


Example commands:
  bash ./aspen -w=/my/output/folder -m=init
  bash ./aspen -w=/my/output/folder -m=dryrun
  bash ./aspen -w=/my/output/folder -m=run
  bash ./aspen -w=/my/output/folder -m=run -c /data/${USER}/.singularity

##########################################################################################

VersionInfo:
  python          : python/3.10
  snakemake       : snakemake
  pipeline_home   : /gpfs/gsfs10/users/CCBR_Pipeliner/Pipelines/ASPEN/dev
  git commit/tag  : 51cb3aee2142ce1226acca97e1e662d54c881a13    v1.0.0-2-g51cb3ae
  aspen_version   : v1.0.0

##########################################################################################

aspen is a Biowulf and FRCE specific wrapper script to the pipeline. Essentially, to run the pipeline the user has to follow 3 steps:

  1. Initialize the output folder:

This can be done using the following command:

% aspen -m=init -w=<path_to_output_folder>

The above command will create config.yaml and samples.tsv in the output folder. Please edit these as per your requirements. You can replace the samples.tsv file in the output folder with the sample manifest created in the previous step outlined above. contrasts.tsv should also be included if running differential ATAC.

  1. Dryrun:

To dry-run the pipeline, you can run the following command after initializing the output folder:

% aspen -m=dryrun -w=<path_to_output_folder>

This should list out the chain of jobs (DAG) that will be submitted to the job scheduler.

  1. RUN!!:

If the dry-run looks as expected, then you can submit the job using:

% aspen -m=run -w=<path_to_output_folder>

This will submit one master job to slurm, which will in turn keep managing the entire pipeline and submit/monitor jobs to the job scheduler as and when required.

3. Genomes

ASPEN supports the following genome versions:

Genome Version Organism
hg38 Human
hg19 Human
mm10 Mouse
mmul10 Macaca mulatta(Rhesus monkey) or rheMac10
bosTau9 Bos taurus(cattle)

4. Disclaimer

This snakemake pipeline is built to run on Biowulf and FRCE. But, as it uses containers for all intermediate steps, the pipeline can be executed on any HPC with minimal edits to the config file.

5. Help

For comments/suggestions/advice please reach out to Vishal Koparde or CCBR_Pipeliner. You can also open a new issue here.

Back to Top

Last update: 2024-01-16