snp-dmx-cancer

This repository contains reproducible code for the Snakemake workflow, benchmark evaluations, and supplementary analyses in our paper evaluating genetic variation-based demultiplexing tools (in particular Vireo) for pooled single-cell RNA sequencing samples in cancer (high-grade serous ovarian cancer (HGSOC) and lung adenocarcinoma).

Paper

For details on the analyses, see our paper: Weber et al. (2021), "Genetic demultiplexing of pooled single-cell RNA-sequencing samples in cancer facilitates effective experimental design", bioRxiv

Contents

Workflow

The Snakemake workflow implements a complete workflow for one dataset and doublets simulation scenario (HGSOC dataset, 20% doublets), using the best-performing set of tools (bulk RNA-seq samples genotype from bcftools, demultiplexing using cellSNP/Vireo) from the benchmark. The workflow is modular, and can be adapted to substitute alternative tools.

Scripts for the Snakemake workflow are saved in workflow/:

• workflow/Snakefile: Snakefile defining the Snakemake workflow
• workflow/scripts/: scripts for the individual steps in the workflow

Benchmark evaluations

Scripts for the benchmark evaluations are saved in benchmarking/:

• benchmarking/Snakefile: Snakefile for the initial steps in the workflow for the benchmark evaluations
• benchmarking/parse_BAM_doublets/: scripts for doublets scenarios
• benchmarking/run_scenarios/: scripts to run tools for each benchmark scenario
• evaluations/: scripts to calculate performance evaluations (precision and recall) and generate plots (precision-recall, runtimes, estimated cost savings)

Supplementary analyses

Scripts for the supplementary analyses are saved in the following directories:

• supplementary_debris/: scripts for additional scenarios including 10%, 20%, or 40% ambient RNA from simulated cell debris or lysed cells
• supplementary_snparray/: scripts for additional scenarios using subset of SNPs from SNP array as genotype reference
• supplementary_healthy/: scripts for additional scenarios for healthy (non-cancer) cell line dataset
• supplementary_downstream/: scripts to run downstream doublet detection tool

Additional scripts

Scripts for additional steps outside the main workflow and benchmark evaluations:

• alternative/: scripts for alternative tools that were not used in the final workflow, which may be useful in the future (e.g. salmon alevin instead of Cell Ranger)
• download_EGA/: script to download data files for lung adenocarcinoma dataset (Kim et al. 2020) from European Genome-phenome Archive (EGA) (requires access to the controlled access data repository)
• download_souporcell/: scripts to download data files for healthy (non-cancer) iPSC cell line dataset from souporcell paper (Heaton et al. 2020) from European Nucleotide Archive (ENA)
• filter_vcf/: script to filter 1000 Genomes Project genotype VCF file to retain only SNPs in 3' untranslated regions (UTRs), for faster runtime
• genotype/: scripts to run different options of tools to generate custom genotype VCF file, including from matched bulk RNA-seq samples (using either bcftools or cellSNP), or directly from single-cell RNA-seq samples (using cellSNP)

Links

• Vireo paper: Huang et al. (2019), Genome Biology
• Vireo documentation
• cellsnp-lite documentation