DecontX snakemake pipeline

Snakemake pipeline that corrects for ambient RNA expression in single-cell RNA sequencing (scRNA-seq) data using DecontX.

Overview

This project contains the code to run a two-phase islet decontamination protocol, as well as configurations and wrapper scripts to run the pipeline on sample data.

Quickstart using Docker

1. Clone the GitHub repo and cd into the repo directory.

# clone the repo
git clone https://github.com/CollinsLabBioComp/islet_decontamination.git

# set code base path
SNK_REPO="$(pwd)/islet_decontamination"
cd ${SNK_REPO}

2. Launch the Docker app and download the Docker image

docker pull letaylor/sc_decontx:latest

3. Run the sample data:

chmod +x ./run_docker.sh
./run_docker.sh

Input data

The expected input data is a folder containing standard 10x outputs. Each folder should contain the following standard folders:

• [sample]/outs/filtered_feature_bc_matrix
• [sample]/outs/raw_feature_bc_matrix

with each [raw/filtered]_feature_bc_matrix folder containing barcodes.tsv.gz, features.tsv.gz, and matrix.mtx.gz. For reference, see the provided sample data.

Configuration

To configure to use your own data:

1. Update workflow/src/threeprime.yaml to change the run ID (name) and specify the sample IDs (samples). note: if your 10x output directory format differs, you may need to update input_dir_basename and input_path_format to match.
2. Place the 10x output folder (containing a minimum of outs/, see here for reference) in the ./data/ folder. Alternatively, you can modify input_dir_base and input_path_format so the base (i.e. data/) points to your parent directory containing all samples.