/dropseqRunner

Lightweight pipeline for processing droplet-based single-cell RNA-seq data

Primary LanguagePython

TLDR

git clone git@github.com:aselewa/dropseqRunner.git
cd dropseqRunner
conda env create -f environment.yaml
conda activate dropRunner

STAR --runThreadN 4 --runMode genomeGenerate --genomeDir $OUTDIR --genomeFastaFiles  $FASTA --sjdbGTFfile $ANNOTATION_GTF 
python dropRunner.py  --R1 path/to/{}.R1.fastq.gz \
                      --R2 path/to/{}.R2.fastq.gz \
                      --indices $OUTDIR \
                      --protocol drop/10x-v3/10x-v2
                      --sample pbmc_v3

If the above give you any trouble, run the demo to ensure everything is installed properly:

make run_test_workflow

Look for a message at the end that tells you whether the demo ran properly or not.

Getting started

dropRunner is a Snakemake-based pipeline for processing single-cell RNA-seq data from the Drop-seq and 10x platform. We utilize STARsolo for alignment and constructing the digital expression matrix. We also supply a detailed report in HTML format that shows the sequencing statistics, as well as read distribution across the genome. The pipeline only works on Linux systems (excluding Windows linux subsystem).

This pipeline is still under active development. If you have issues, please report them via GitHub.

Setting up conda

You may skip this if you already have conda installed and configured.

miniconda3 is a light version of Anaconda. To install on 64Linux, do the following:

wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
bash Miniconda3-latest-Linux-x86_64.sh

Once it is done, initialize conda so it is in your path:

conda init bash
source ~/.bashrc

Check conda works:

conda --version

0. Set up and activate environment

Use the provided environment.yaml file to set up the conda environment.

git clone git@github.com:aselewa/dropseqRunner.git
cd dropseqRunner
conda env create -f environment.yaml

This may take some time depending on your environment. A fresh conda installation should take about 10-15 minutes.

Activate the environment before running the next steps

conda activate dropRunner

1. Make reference genome indices

Use STAR to create indices for your reference genome of interest. You will need two things:

  • fasta file of reference genome
  • reference genome GTF annotations

You can get both of these from GENCODE for humans.

STAR --runThreadN 4 --runMode genomeGenerate --genomeDir $OUTDIR --genomeFastaFiles  $FASTA --sjdbGTFfile $ANNOTATION_GTF

2. Run the pipeline

Use dropRunner.py on your fastq files to generate count matrices. Use the protocol parameter to specify drop, 10x-v2, or 10x-v3. The last two are version 2 and version 3 10x platforms.

python dropRunner.py  --R1 path/to/{}.R1.fastq.gz
                      --R2 path/to/{}.R2.fastq.gz
                      --indices $OUTDIR
                      --cluster
                      --sample my_example_project

Note 1: You can supply multiple R1s and R2s by passing a comma-delimited list. I find this bash command useful:

R1=$(ls *.R1.fastq.gz | paste -sd,)

3. Output

There are two pieces of information that most users will need:

  • html reports
  • count matrices

The html report is in reports/. The count matrices are in output/{project_name}_Solo.out. There are two types of count matrices: filtered and raw. The raw matrix contains all valid barcodes, while filtered contains only barcodes with a certain number of UMI. This threshold is determined by STARsolo using a hueristic approach.

Please report any issues you run into via GitHub.