Making ENCODE ChIP-seq pipeline work on MSI

Installation

It's worth doing this interactively, since it will take a while. This repo is identical to the main repo except for some changes to the install script to make a local conda install with the --prefix command. I recommend installing this repo to your groups shared directory, that way it's easily usable across your lab. Furthermore, this install will add over 100K files to your groupquota, so you don't want it replicated a bunch!

# Request an interactive session

srun --nodes=1 --ntasks-per-node=10 --cpus-per-task=1 --time=03:00:00 --mem=10GB --account=<group_name> --partition=interactive --pty bash

# Load Python3 module with mamba
module load python3/3.8.3_anaconda2020.07_mamba

cd /home/<group_name>/shared 
git clone https://github.com/acherman/chip-seq-pipeline2.git
cd chip-seq-pipeline2/

# This might take a while!
bash scripts/install_conda_env.sh mamba

Activation and de-activation

module load python3/3.8.3_anaconda2020.07_mamba

# Use this if you don't have your shell configured or haven't run conda (either ever, or in your current session)
source activate /home/group/me/software/chip-seq-pipeline2/encode-chip-seq-pipeline

# Otherwise, use this
conda activate /home/group/me/software/chip-seq-pipeline2/encode-chip-seq-pipeline

# De-activation
conda deactivate

Dealing with bowtie2 error

There seems to be a problem with tbb library access in bowtie2. While not exactly the same issue, I believe the solution can be found here. Essentially, we need to install a particular version of tbb.

module load python3/3.8.3_anaconda2020.07_mamba

source activate /home/group/me/software/chip-seq-pipeline2/encode-chip-seq-pipeline

conda install tbb=2020.2

This should hopefully do the trick!

TBD

I do not know how this software is actually run! Once this is figured out, I'll add a job script example. If it's easier to run interactively, it's easiest to modify the srun command above for more memory, CPUs, and walltime.