Structural variant calling from single-cell Strand-seq data - summarized in a single Snakemake pipeline.

Overview of this workflow

This workflow uses Snakemake to execute all steps of MosaiCatcher in order. The starting point are single-cell BAM files from Strand-seq experiments and the final output are SV predictions in a tabular format as well as in a graphical representation. To get to this point, the workflow goes through the following steps:

1. Binning of sequencing reads in genomic windows of 100kb via mosaicatcher
2. Normalization of coverage with respect to a reference sample (included)
3. Strand state detection (included)
4. Haplotype resolution via StrandPhaseR
5. Multi-variate segmentation of cells (mosaicatcher)
6. Bayesian classification of segmentation to find SVs using mosaiClassifier (included)
7. Visualization of results using custom R plots (included)

System requirements

This workflow is meant to be run in a Unix-based operating system (tested on Ubuntu 18.04).

However, several Docker images are available for portability (see Installation).

Minimum system requirements vary based on the use case. The whole pipeline can be run in a notebook (8GB RAM, 4 cores, 3GHz CPUs), but we highly recommend running it in a server environment with 32+GB RAM and 24 cores.

Installation

Choose one of many ways to install and run this workflow, from easiest to use to most flexible:

1. Run a complete example via Docker [Demo data set]

   • Instructions here
   • Requires Docker (tested in version 18.09), no further setup required
   • In case of testing in a Docker Desktop, please make sure the setting of resources (Memory: 3.0 GiB, Swap: 4.0GiB), this can be checked/changed via 'Preferences -> Advanced' in your Docker Desktop software

2. Run your own data set via Docker

   • Instructions here
   • Requires Docker (tested in version 18.09) and a manual setup of data (Setup)

3. Run Snakemake together with a Singularity image

   • Instructions here
   • Requires Snakemake and Singularity.
   • Add your data and configuration as described below (Setup)
   • More flexible than Docker since snakemake is run on your system (not within the container)

4. Install software using Bioconda

   • Installation instructions here
   • Configure Snake.config.json to match your software installation
   • Add your data and configuration as described below (Setup)

Setup

1. Download this pipeline

   git clone https://github.com/friendsofstrandseq/pipeline
   cd pipeline
   

2. Add your single-cell data

   Create a subdirectory bam/sampleName/. Your Strand-seq BAM files of this sample go into two folders:

   • all/for the total set of BAM files
   • selected/ for the subset of successful Strand-seq libraries (possibly hard-linked to all/)

   It is important to follow these rules for single-cell data

   • One BAM file per cell
   • Sorted and indexed
   • Timestamp of index files must be newer than of the BAM files
   • Each BAM file must contain a read group (@RG) with a common sample name (SM), which must match the folder name (sampleName above)

3. Adapt the config file

   Set options to describe your data in Snake.config.json. If you use Singularity, please use Snake.config-singularity.json instead.

   Here is a digest of the relevant variables:

   • reference: The path to the reference genome (FASTA file). Must be indexed (FAI)
   • chromosomes: Specify which chromosomes should be analyzed. By default chr1 - chr22 and chrX
   • R_reference The version of the reference genome being used by R scripts. You will need to install the respective R package (e.g. BSgenome.Hsapiens.UCSC.hg38) on your system to be able to get this file. The reference version should match the one in reference. Note that the Singularity/Docker image only ship with BSgenome.Hsapiens.UCSC.hg38.
   • paired_end: Specifies whether you are using paired-end reads or not (default: true)
   • snv_calls: SNV call set for your sample - see below. Must be vcf.gz and indexed via tabix
   • snv_sites_to_genotype: SNV positions to be genotyped - see below. Must be vcf.gz and indexed via tabix

SNV calls for haplotype separation

The MosaiCatcher pipeline requires a set of genotyped single nucleotide variants (SNVs) to distinguish haplotypes, including the assignment of individually sequenced strands of a chromosome to a certain chromosome-length haplotype.

Depending on which prior information is available, the workflow will try to

• Phase SNVs using StrandPhaseR (when given a set of genotyped SNV calls)
• Genotype & phase SNVs in your data (when given potential SNV sites)
• Call, genotype & phase SNVs on your sample (when given no SNV information)

To provide a list of SNV sites, set snv_sites_to_genotype in the config file; to provide genotyped final SNV calls, set snv_calls. Must be set per sample:

"snv_calls"     : {
	"NA12878" : "path/to/snp/calls.vcf.gz"
},

Make sure the files are tabix-indexed.

References

For information on Strand-seq see

Falconer E et al., 2012 (doi: 10.1038/nmeth.2206)