Snakemake workflow to align SARS-CoV-2 paired-end sequencing reads to NCBI reference sequence NC_045512.2.

The workflow preprocesses and maps reads to NC_045512.2 Wuhan seafood market pneumonia virus isolate Wuhan-Hu-1 using Bbmap, calls variations with Lofreq and generates consensus fasta sequences with sites with zero coverage masked.

This workflow was used to analyse SARS-CoV-2 sequences under the KoroGeno-EST project.

Main outputs:

• results/consensus_masked.fa -- multi FASTA file with generated consensus sequences.
• results/snpsift.csv -- csv file with filtered variant positions that were used to generate consensus sequences.
• results/multiqc.html -- aggregated QC report in html format.

• Download and install miniconda3: https://docs.conda.io/en/latest/miniconda.html.
• Follow instructions to install snakemake: https://snakemake.readthedocs.io/en/stable/getting_started/installation.html

• Create a new working directory e.g. covidseq and clone this repository to working directory.

mkdir covidseq
git clone https://github.com/avilab/sarscov2-variation.git covidseq
cd covidseq

Sample configuration uses now portable encapsulated project (PEP) definition.

• Create/Edit config/pep.yaml and config/samples.csv files. Please see test directory for examples.

Example of samples.csv file with paired reads in two separate files:

In case of interleaved fastq files, following samples.tsv can be used:

• Update config/pep.yaml: "read1" (and "read2", "batch") in sources definition matches column(s) in your samples.csv file, if you have other column names, adjust this variable accordingly in pep.yaml. Please see PEP specification for further options to customise pep config file to your needs.

pep_version: 2.0.0
sample_table: "samples.csv" OR "samples.tsv"
sample_modifiers:
  append:
    fq1: r1
    fq2: r2
  derive:
    attributes: [fq1, fq2]
    sources:
      r1: "/path/to/reads/{batch}/{read1}"
      r2: "/path/to/reads/{batch}/{read2}"

In case of interleaved reads:

pep_version: 2.0.0
sample_table: "samples.csv" OR "samples.tsv"
sample_modifiers:
  append:
    fq: r1
  derive:
    attributes: [fq]
    sources:
      r1: "/path/to/reads/{batch}/{read1}"

(Optional) Human and rRNA sequence databases for FastQ Screen. In absence of databases fastq screen rule will be skipped.

Human genomic sequence database is used to estimate and remove human sequences from analysis. Run scripts/download_masked_human_hg19.sh to download masked human reference genome to filter out reads mapping to the human genome. Move hg19_main_mask_ribo_animal_allplant_allfungus.fa.gz file in your system where you store databases. hg19_main_mask_ribo_animal_allplant_allfungus.fa.gz file was indexed using bwa index command. Setup environment variable "REF_GENOME_HUMAN_MASKED" pointing to this file or edit "HOST_GENOME" variable in Snakefile.

Silva rRNA database is used to estimate and remove rRNA contamination. Database files can be downloaded from https://www.arb-silva.de/fileadmin/silva_databases/release_138/Exports/SILVA_138_SSURef_NR99_tax_silva.fasta.gz and https://www.arb-silva.de/fileadmin/silva_databases/release_132/Exports/SILVA_132_LSURef_tax_silva.fasta.gz and moved to systems' databases folder. LSU and SSU fasta files were concatenated and indexed using bwa index command. Setup environment variable "SILVA_DB" pointing to this file or edit "RRNA_DB" variable in Snakefile.

Analyse sequences in the test folder:

snakemake --use-conda -d .tests/integration -j 1

Dry run:

snakemake --use-conda -n

Analyse sequences:

snakemake --use-conda -j

For all possible snakemake command line options please refer to snakemake tutorial https://snakemake.readthedocs.io/en/stable/executing/cli.html.

This workflow can be run on a contemporary PC/laptop (e.g. i5/16G) with sufficient HD space to accomodate sequening runs.

Workflow graph can be generated:

snakemake --dag -d .tests/integration | dot -Tsvg > images/rulegraph.svg