A snakemake pipeline for mapping reads from various types of plate or droplet-based single-cell platforms, for crossover/haplotyping analysis using snco. The main snco tool repo is available here.
Several droplet-based single-cell modalities are currently supported, including 10x 3' RNA v3 and v4, as well as BD Rhapsody 3' RNA and 10x ATAC. For these datasets one or more pairs of fastq files containing reads from all barcodes should be supplied.
Plate-based methods like Takara PicoPlex WGA or analysis of whole-genome resequencing datasets for backcross/F2 populations (see Rowan et al. 2015 for an example) are supported using individual input fastq files for each barcode and/or individual.
git clone https://github.com/schneebergerlab/snco_mapping_pipeline.git
cd snco_mapping_pipeline
pip install -e .
snco_pipeline init
This creates a snco_mapping_config.yaml in your current directory, prompting you to fill in some of the required fields. The pipeline run requires a directory structure specified by the config file. The annotation_dir directory should contain the reference genomes and annotations, plus optional predefined variants, in fasta, gtf, and vcf formats respectively. The raw_data_dir directory should contain the input fastq files. The results_dir directory will be populated by the aligned data and haplotyping results during the pipeline run.
snco_pipeline run
The snakefiles directing the pipeline structure are located with the installation of the tool and are not present in the output directory. This is to make it easier to create multiple independent runs of the pipeline without copying/symlinking snakefiles.
The snco_pipeline cli is a fairly straightforward wrapper of the snakemake cli. You can supply any Snakemake arguments (e.g. --profile, --use-conda, etc.) which will be passed directly onwards to snakemake:
snco_pipeline run --cores 4 --use-conda --rerun-incomplete
The only arguments that should be avoided are --snakefile and --configfile since these are managed by the cli.
The pipeline itself has few requirements - just snakemake, jinja2, and click. None of the tools used for the analysis are installed by the setup.py script itself, but are instead managed by snakemake using conda. The default environment yamls used by snakemake are stored in the subdirectory snco_mapping_pipeline/snco_pipeline/pipeline/env_yamls/ of this repo. If you have problems with the conda installations of these environments, alternative yamls or existing conda environment names can be provided in the conda_envs section of the snco_mapping_config.yaml file.