peako 0.0.3

A package for discovering motifs in ChIP-seq datasets with knockout controls

PeaKO

What is peaKO?

PeaKO discovers motifs in ChIP-seq datasets with knockout controls. PeaKO takes in paired wild-type/knockout BAM files in addition to several reference files, as input. It returns a file of ranked motifs (see our paper for more details).

Quick start

Dependencies

1. Conda (Miniconda or Anaconda)
2. MEME Suite version 4.12.0 with our CentriMo binary* (see below)

Installation

1. Download peaKO's environment file.
2. Open a terminal and run conda env create -f peako-env.yml in your Downloads directory. This will create a Conda environment called "peako".
3. Run conda activate peako or source activate peako to activate this environment.
4. Install peaKO from PyPI by running python3 -m pip install peako.
5. You can test that this worked by running peako --help.

Instructions for our modified CentriMo binary

*Our modified CentriMo application will be incorporated in MEME Suite's next major release. Until then, you may install MEME Suite from source and replace its binary with our own to use peaKO.

1. Download MEME distribution 4.12.0 from the MEME Suite Download page.
2. Follow the "Quick Install" steps on the MEME Suite Installation page up until make install.
3. After running make install, replace $HOME/meme/bin/centrimo with our modified CentriMo binary.
4. Make sure that $HOME/meme/bin is located on your $PATH. You should now be able to call centrimo --help.

Usage

PeaKO uses Snakemake, which is a workflow management system. You can run peaKO either locally or on a compute cluster using the Slurm job scheduling system. To run on Slurm, you must create your own cluster.config file (template) and provide it to peaKO via --sm-cluster-config.

Each step of the workflow either inherits from the main activated Conda environment ("peako") or uses its own separate environment. If you are working on a compute cluster, run peaKO first with --sm-build-envs on a node with internet access to create these additional Conda environments. Then, you can run it on the cluster without internet, providing a Slurm configuration file (see above).

After activating peaKO's Conda environment (conda activate peako or source activate peako), you can run peako as follows:

peako <outdir> <wt-bam> <ko-bam> <organism> <chr-sizes> <trf-masked-genome> <motif-database> [options]

There are 7 required arguments. Please provide full paths for files and directories.

• outdir: output directory (please make sure this already exists); all output directories and files will be created here
• wt-bam: wild-type sample BAM file
• ko-bam: knockout sample BAM file
• organism: name of organism (must be either mouse or human)
• chr-sizes: chromosome sizes file of reference genome (TXT)
• trf-masked-genome: TRF masked reference genome file (FASTA)
• motif-database: JASPAR motif database (MEME)

Here are the optional arguments:

General:

• -h or --help: access the help message and exit
• -V or --version: show the program's version and exit

PeaKO submodule:

• -j <JASPAR_ID>: transcription factor motif JASPAR identifier (e.g. MA0083.3)
• -m <MOTIF>: transcription factor motif common name (e.g. SRF)
• --extra: output all intermediate peaKO files for plotting
• --pickle: use pickled peaKO dictionaries from previous run

Snakemake:

• --sm-build-envs: build conda environments for workflow and exit (requires internet connection)
• --sm-cluster-config: snakemake cluster configuration file (JSON)

Output

Currently, peaKO generates output directories and files for each step. These can all be found under your provided outdir directory. PeaKO's main output file is <outdir>/peako_out/peaKO-rankings.txt, which contains a ranked list of motifs.

Additional resources

Source code is available at: https://github.com/hoffmangroup/peako.

We have deposited the current version of the code, example HTML and TXT CentriMo outputs, and a modified CentriMo binary on Zenodo.

Citation

If you found peaKO useful, please cite:

Denisko D, Viner C, Hoffman MM. Motif elucidation in ChIP-seq datasets with a knockout control. BioRxiv 10.1101/721720 [Preprint]. 2019. Available from: https://doi.org/10.1101/721720