ATAC-seq Analysis

Prerequisites

Your system MUST have Homebrew or Linuxbrew installed.

Linuxbrew: http://linuxbrew.sh/
Homebrew: https://brew.sh/

You must also have Java 1.5 or later installed (most Linux and Mac boxes do.)

The script install-dependencies.sh assumes you have those dependencies available and will fail without them.

Setup

Once you've cloned or downloaded this repository, drop all its files into the root directory containing your files needing analysis.
Go to https://conda.io/miniconda.html , download, and run the installer appropriate for your platform. (The OSX installer is included in this repo).

That's all!

Configuration

Open a terminal at the root of this repository and run

sudo chmod +x *.sh
./install-dependencies.sh

This will configure your system for the rest of the project. This step may take up to an hour or two to complete.

Problems on OSX?

If you're on OSX some versions intermittently complain about the file ending of scripts. If it does, instead run

mv *.sh *.command
sudo chmod +x *.command
./install-dependencies.command

to configure your dependencies.

Generating Bedfiles

To convert your gzipped or unzipped SAM files to binary, peak, index, and bedfiles, run:

python ./samlooper.py

This script has a single configuration variable: at the top, set DO_FOOTPRINTING to True if you want to analyse footprints and profiles, and False otherwise.

Note this is VERY SLOW. This will take 1-2 hours per file in total to process fully (unzip, convert to binary, index, quality filter, peak find, bed creation, footprint, and profile).

Processing Bedfiles

TODO

Annotating Bedfiles

TODO

Running Counts

Be sure you've generated your bedfiles -- they also generate the sorted q2 *.bam files you need to get this part going.

You'll need to edit do-htseq-counts.py. Near the top of the file, after the docstring, you'll find a python dictionary datatype object named bamPrettyMap.

Edit this file to set the labels you desire as the (unique) dictionary key, and the *.sorted.q2.bam filename as the value, for example:

bamPrettyMap = {
    "LYR_CD_CON2": "AT-A-11_S9_BowtieOut.sorted.q2.bam",
    "LYR_CD_CON3": "AT-B-38_S22_BowtieOut.sorted.q2.bam",
    "LYR_CD_DR1": "AT-A-16_S1_BowtieOut.sorted.q2.bam"
}

If your key has a underscore-separated prefix, it'll only match those samples againstgff3 files that contain that (case-sensitive) substring; eg, in the example above, only *.gff3 containing LYR will be used for matching. If there is no underscore-separated prefix, no species matching will be used.

By default, the file also has a configuration value overwriteExistingCounts set to False. Set this value to True to regenerate everything each time you run this script.

Then, once you've done that, simply run

python ./do-htseq-counts.py

Counts will be output as both csv and txt.

tigerhawkvok/atacseq-helper