Skip to content

Latest commit

 

History

History
48 lines (31 loc) · 3.15 KB

README.md

File metadata and controls

48 lines (31 loc) · 3.15 KB
Raw

segway

CircleCI Code style: black MIT License

Description

A pipeline to run Segway and Segtools to generate chromatin segmentation models and tracks.

Installation

  1. Git clone this pipeline.

    $ git clone https://github.com/ENCODE-DCC/segway-pipeline.git

  2. Install Caper, requires java > 1.8 and python > 3.4.1 . Caper is a python wrapper for Cromwell.

    $ pip install caper  # use pip3 if it doesn't work

  3. Follow Caper's README carefully to configure it for your platform (local, cloud, cluster, etc.)

    IMPORTANT: Configure your Caper configuration file ~/.caper/default.conf correctly for your platform.

Usage

To run the test case, cd to the segway-pipeline repository cloned in the installation, and run the following:

$ caper run segway.wdl -i tests/functional/json/test_segway_full.json --docker

Input Data

In the scripts directory there is a script to generate lists of input. It takes the ENCODE accession of a reference epigenome as an argument, finds the appropriate files to use as input to the model, and generates an input JSON for the pipeline. It will filter out control experiments and non-continuous datasets like WGBS and identify bigWig files on GRCh38 for each non-control experiment, preferring pooled files if the experiment is replicated. For ChIP-seq and ATAC-seq, bigWigs with the output type fold change over control will be selected. For DNase datasets, the read-depth normalized signal bigWig from the replicate with the greatest number of mapped reads after filtering will be selected.

To install dependencies for the scripts, make sure you have Python >= 3.5 and run pip install -r requirements-scripts.txt

An example usage is given below. The values for --chrom-sizes and --annotation-gtf are IDs for file objects at the portal, e.g. https://www.encodeproject.org/files/GRCh38_EBV.chrom.sizes . The assays to skip specified by --skip-assays should be quoted to avoid being consumed as separate arguments if there are spaces in them. The --chip-targets correspond to an experiment's target.label property, and can be either histone or TF targets. An error will be raised if no matching targets were found for the given reference epigenome.

python scripts/make_input_jsons_from_portal.py --chrom-sizes GRCh38_EBV.chrom.sizes --annotation-gtf gencode.v29.primary_assembly.annotation_UCSC_names  --skip-assays "TF ChIP-seq" --chip-targets H3K4me3 H3K27ac -a ENCSR867OGI

Development

See the developer docs for more details on running tests and developing this pipeline.