Learning Shared Chromatin Landscapes and Joint De-Noising of Histone Modification Assays with DecoDen

DecoDen uses replicates and multi-histone ChIP-Seq experiments for a target cell type to learn and remove shared biases from fragmentation, PCR amplification and sequence mappability.

NEW! Check out the Jupyter notebook tutorial, or read on.

Installation

The installation of DecoDen is currently offered as a Poetry project while in development. The procedure proposed requires a local installation of git and a C compiler. We recommend the use of Conda to create a suitable environment, with a command such as conda create -n decoden python>=3.9. After the activation of the environment (conda activate decoden), follow these steps:

1. Install Poetry
2. Clone the repository and install with poetry by running these commands

# Clone the repository
git clone git@github.com:ntanmayee/decoden.git
cd decoden

# Install the external dependencies and DecoDen
conda config --add channels bioconda
conda config --add channels conda-forge
conda install poetry pysam=0.22 zlib bioconda::ucsc-bedgraphtobigwig

# If there is no C compiler installed include also the following command
conda install c-compiler

poetry install

Quick Start

Input data

Running decoden requires two inputs:

1. Aligned reads in .bam format from ChIP-Seq experiments
2. Sample annotation file in .csv format

Auto-generate a sample annotation file

To generate a skeleton sample annotation file, run -

decoden create_csv 

This will create samples.csv in your current directory. Edit this file and fill in the columns with appropriate information. There are more details here.

Run DecoDen

Run the DecoDen pipeline with default parameters -

decoden run -i samples.csv -o output_directory -gs genome-size

Results

Results are stored in feather format in HSR_results.ftr. For each assay type, a separate bigwig file is saved in bw directory.

Detailed Usage Guidelines

The following commands are available in DecoDen. Please click on the links to know more about them.

Command	Description
create_csv	Create a skeleton sample annotation file
run	Run the full DecoDen pipeline to preprocess end denoise BAM/BED files
preprocess	Pre-process BAM/BED data to be in the correct format for running DecoDen
denoise	Run the denoising step of DecoDen on suitably preprocessed data
detect	Detect peaks in the processed DecoDen signals

There is more helpful information in the wiki.

Bug Reports and Suggestions for Improvement

Please raise an issue if you find bugs or if you have any suggestions for improvement.

Funding

This project has received funding from the European Union's Framework Programme for Research and Innovation Horizon 2020 (2014-2020) under the Marie Skłodowska-Curie Grant Agreement No. 813533-MSCA-ITN-2018