atpoint/sc_preprocess repository overview

⁠sc_preprocess

A Nextflow pipeline for preprocessing of 10X scRNA-seq data using Alevin.

⁠Workflow

• download genome, transcriptome and GTF reference files from GENCODE version M25⁠.

• create an index of spliced- and unspliced transcript sequences based on the reference annotations, following the [Alevin Velocity tutorial(https://combine-lab.github.io/alevin-tutorial/2020/alevin-velocity/⁠)

• create a merged of the exon+intron transcriptome with the mouse genome into a gentrome and index it with salmon

• perform cell barcode detection, UMI deduplication and read quantification with Alevin⁠

• read quantifications into R with tximeta, split into spliced- and unspliced counts and save as MatrixMarket (mtx) to disk for easy distribution of data

⁠Usage

As the workflow is hosted on GitHub it can be pulled directly via nextflow. On the HPC we use:

NXF_VER=21.04.1 nextflow run atpoint/sc_preprocess -r <commit_sha> -profile singularity,slurm --fastq 'path/to/*_{1,2}.fastq.gz'  

Details:

• --fastq: path to fastq file pairs, e.g. --fastq path/to/*_{1,2}.fastq.gz

• -profile: choose a container profile, -profile singularity or -profile docker (see #software) This repository contains an environment.yml to create a conda environment. This is not available directly via the workflow so one would need to manually create an environment, e.g. mamba create --name sc_preprocess --file environment.yml and then run the workflow locally.
  For testing one can add -profile test. With -profile slurm nextflow will submit jobs via the SLURM scheduler.

• --idx_label & --quant_label allocate resources:
  => for the indexing there is --idx_label idx_small/idx_big which allocates 4 or 36 cores and 15 or 40GB of RAM.
  => for the quantification there is --quant_label quant_small/quant_big which allocates 3 or 36 cores with 6 or 50GB RAM.
  => the defaults are the big labels intended for HPC use.

We use a naming convention of the fastq files like:

• basename_1.fastq.gz for R1 (CB/UMI)

• basename_2.fastq.gz for R2 (cDNA)

⁠Software

A Docker container is available at Docker Hub⁠.
When using the Docker or Singularity profile Nextflow will take care of pulling it.

⁠Citations

• nf-core project⁠

• Ewels et al (2020) Nature Biotechnology volume 38, pages 276–278⁠

• Nextflow Docs⁠

• Seqera Training⁠

• https://github.com/nextflow-io/rnaseq-nf⁠

• Merkel, D (2014). Docker: lightweight linux containers for consistent development and deployment. Linux Journal⁠

• Kurtzer et al (2017) Singularity: Scientific containers for mobility of compute. PLoS ONE⁠

• Srivastava et al (2019) Alevin efficiently estimates accurate gene abundances from dscRNA-seq data. Genome Biology⁠

• Soneson et al (2021) Preprocessing choices affect RNA velocity results for droplet scRNA-seq data. PLoS Computational Biology⁠

• Love et al (2020) Tximeta: Reference sequence checksums for provenance identification in RNA-seq⁠