nf-ncov-voc

Introduction

nf-ncov-voc is a bioinformatics workflow developed to process viral genomes and integrate the contextual data. The workflow was intially designed for processing SARS-CoV-2 for COVID-19 pandemic response and has been later adapted for more priority viruses e.g., Mpox and Influenza. The workflow is developed in a modular structure with several modules and sub-workflows leveraged from nf-core. These modules and sub-workflows are assembled in a plug-n-play manner based on the data and viral charatceristics. Each virus supported by the workflow has its own workflow file that directs the assembly of sub-workflows and modules.

The workflow is built using Nextflow- DSL2, a workflow tool to run tasks across multiple compute infrastructures in a very portable manner. It can use conda/Docker/Singularity containers making installation trivial and results highly reproducible.

nf-ncov-voc workflow can be used in combination with an interactive visualization tool VIRUS-MVP or as a stand-alone high-throughput analysis tool to produce mutation profiles and surveillance reports.

A detailed structure and each module of the workflow is presented below in the dataflow diagram

nf-ncov-voc Dataflow

Functional Annotation

nf-ncov-voc offers a unique opportunity to integrate the contextual data with genomics data. Variant Called File (VCF) generated for each group or sample is then converted to a Genome Variant File (GVF) to integrate the functions associated to different mutations. For more information of how the functions are curated and structured see the dedicated repository Pokay

nf-ncov-voc with the help of functional data in Pokay, produces surveillance reports that are developed in collaboration with the Public Health partners and offers a high-level yet comprehensive report on each mutation its associated functions in literature.

Input data

As an input, nf-ncov-voc can accept different formats, Whole Genome Sequences (WGS) in FASTA format with a Metadata file in TSV format; paired-end short read sequences in FASTQ format with a Metadata file in TSV format. Additionally, the input can also be VCF file that contains variants called.

Grouping data

Quality control

Variant Calling

Sequences in pre-processing stage are filtered using Metadata variables, quality filtered and assigned lineages. Sequences assigned as VOCs, VOIs and VUMs are then mapped to SARS-CoV-2 genome, variant called and normalized in Genomic Analysis (Variant Calling) module. Mutations called are then annotated in several stages including flagging the potential contaminated sites, mutation annotation, genomic feature annotation, mature peptide annotation and finally respective biological functional impact using the manually curated effort Pokay. (lead by Paul Gordon @nodrogluap). Finally, in the surveillance module, these functional profiles are summarized using functional indicators to highlight key functions and mutations responsible for them for e.g. P618H role in convalescent plasma escape.

Pre-Processing

This module offers two ways to get lineage information for each genome in FASTA file and listed respectively in Metadata file unless a column pango_lineage is already available in which case both options can be skipped. First option is to use PANGOLIN to assign lineages and merge the metadata with pangolin report. This step can be skipped by passing --skip_pangolin. The second option is to map input metadata to GISAID metadata file (which can be provided by --gisaid_metadata parameter) if the genomes are available in GISAID. This option is faster and computationally less expensive, though limits to only genomes available in GISAID. This option can be skipped by using --skip_mapping.

Genomic Analysis

This module currently supports two different modes - "reference" & "user" which can be passed with --mode reference or --mode user. By default, --mode reference is activated which allows you to build a reference library for each lineage and subsequently each variant for comparative analysis. This mode can take FASTA file with multiple genomes (recommended & default) or single genome with a metadata file that should have one column atleast (pango_lineage) as minimal metadata (see Workflow Summary for detailed options). The workflow has numerous options for several steps. For example, in mode --reference user can use BWAMEM using --bwa instead of MINIMAP2 (default) for mapping consensus sequences to reference genome. Similarly, ivar with parameter --ivar for variant calling instead of freebayes (default) option. The user mode (--mode user) is by default active when using interactive visualization through COVID-MVP where a user can upload GVF file for comparative analysis against the reference data. Uploaded dataset can be a FASTA file or variant called VCF file.

Functional Annotation

In this module, the variant called VCF file for each lineage is converted into a GVF (Genomic Variant Format) file and annotated with functional information using Pokay. GVF is a variant of GFF3 format that is standardized for describing genomic mutations; it is used here because it can describe mutations across multiple rows, and because the "#attributes" column can store information in custom key-value pairs. The key-value pairs added at this stage include for each mutation: VOC/VOI status, clade-defining status (for reference lineages), and functional annotations parsed using vcf2gvf.py file written in python.

Surveillance Reports

Different GVF files for the same variant are then collated and summarized into a TSV file that contains mutation prevalence, profile and functional impact. Further TSV file is also summarized as a more human friendly and impactful surveillance report in a PDF format. Relevant/important indicators can be specified in the tsv file. This feature of surveillance reports can be used to identify new clusters, important mutations, and track their transmission and prevalence trends. However, if not required, this step can be skipped using --skip_surveillance. An example of surveillance file for Omicron variant using VirusSeq Data Portal is available in Docs

See the parameters docs for all available options when running the workflow.

** Further developments will continue to adapt nf-ncov-voc to other viruses in near furture. **

Usage

1. Install Nextflow (>=21.04.0)

2. Install any of Docker, Singularity or Conda for full pipeline reproducibility see recipes

3. Download the pipeline and run with help for detailed parameter options:

   nextflow run nf-ncov-voc/main.nf --help

   N E X T F L O W  ~  version 21.04.3
   Launching `main.nf` [berserk_austin] - revision: 93ccc86071
   
   Usage:
    nextflow run main.nf -profile [singularity | docker | conda) --prefix [prefix] --mode [reference | user]  [workflow-options]
   
   Description:
    Variant Calling workflow for SARS-CoV-2 Variant of Concern (VOC) and
    Variant of Interest (VOI) consensus sequences to generate data
    for Visualization. All options set via CLI can be set in conf
    directory
   
   Nextflow arguments (single DASH):
    -profile                  Allowed values: conda & singularity
   
   Mandatory workflow arguments (mutually exclusive):
    --prefix                  A (unique) string prefix for output directory for each run.
    --mode                    A flag for user uploaded data through visualization app or
                              high-throughput analyses (reference | user) (Default: reference)
   
   Optional:
   
   Input options:
    --seq                     Input SARS-CoV-2 genomes or consensus sequences
                              (.fasta file)
    --meta                    Input Metadata file of SARS-CoV-2 genomes or consensus sequences
                              (.tsv file)
    --userfile                Specify userfile
                              (fasta | vcf) (Default: None)
    --gisaid_metadata         If lineage assignment is preferred by mapping metadata to GISAID
                              metadata file, provide the metadata file (.tsv file)
    --variants                Provide a variants file
                              (.tsv) (Default: /Users/au572806/GitHub/nf-ncov-voc/assets/ncov_variants/variants_who.tsv)
    --outdir                  Output directory
                              (Default: /Users/au572806/GitHub/nf-ncov-voc/results)
    --gff                     Path to annotation gff for variant consequence calling and typing.
                              (Default: /Users/au572806/GitHub/nf-ncov-voc/assets/ncov_genomeFeatures/MN908947.3.gff3)
    --ref                     Path to SARS-CoV-2 reference fasta file
                              (Default: /Users/au572806/GitHub/nf-ncov-voc/assets/ncov_refdb/*)
    --bwa_index               Path to BWA index files
                              (Default: /Users/au572806/GitHub/nf-ncov-voc/assets/ncov_refdb/*)
   
   Selection options:
   
    --ivar                    Run the iVar workflow instead of Freebayes(default)
    --bwamem                  Run the BWA workflow instead of MiniMap2(default)
    --skip_pangolin           Skip PANGOLIN. Can be used if metadata already have lineage
                              information or mapping is preferred method
    --skip_mapping            Skip Mapping. Can be used if metadata already have lineage
                              information or PANGOLIN is preferred method
   
   Preprocessing options:
    --startdate               Start date (Submission date) to extract dataset
                              (yyyy-mm-dd) (Default: "2020-01-01")
    --enddate                 Start date (Submission date) to extract dataset
                              (yyyy-mm-dd) (Default: "2022-12-31")
   
   Genomic Analysis parameters:
   
    BBMAP
    --maxns                   Max number of Ns allowed in the sequence in qc process
    --minlength               Minimun length of sequence required for sequences
                              to pass qc filtration. Sequence less than minlength
                              are not taken further
   
    IVAR/FREEBAYES
    --ploidy                  Ploidy (Default: 1)
    --mpileupDepth            Mpileup depth (Default: unlimited)
    --var_FreqThreshold       Variant Calling frequency threshold for consensus variant
                              (Default: 0.75)
    --var_MaxDepth            Maximum reads per input file depth to call variant
                              (mpileup -d, Default: 0)
    --var_MinDepth            Minimum coverage depth to call variant
                              (ivar variants -m, freebayes -u Default: 10)
    --var_MinFreqThreshold    Minimum frequency threshold to call variant
                              (ivar variants -t, Default: 0.25)
    --varMinVariantQuality    Minimum mapQ to call variant
                              (ivar variants -q, Default: 20)
   
   Surveillance parameters:
    --virusseq                True/False (Default: False). If your data is from
                              VirusSeq Data Portal (Canada's Nation COVID-19
                              genomics data portal).
                              Passing this argument adds an acknowledgment
                              statement to the surveillance report.
                              see https://virusseq-dataportal.ca/acknowledgements

4. Start running your own analysis!

   • Typical command for reference mode when Metadata File don't have lineage information:

     nextflow nf-ncov-voc/main.nf \
         -profile <conda, singularity, docker> \
         --prefix <testing> \
         --mode reference \
         --startdate <2020-01-01> \
         --enddate <2020-01-01> \
         --seq <Sequence File> \
         --meta <Metadata File> \
         --skip_mapping \
         --outdir <Output Dir>

   • Typical command for reference mode when Metadata File already have lineage information:

     nextflow nf-ncov-voc/main.nf \
         -profile <conda, singularity, docker> \
         --prefix <testing> \
         --mode reference \
         --startdate <2020-01-01> \
         --enddate <2020-01-01> \
         --seq <Sequence File> \
         --meta <Metadata File> \
         --skip_mapping \
         --skip_pangolin \
         --outdir <Output Dir>

   • An executable Python script called functional_annotation.py has been provided if you would like to update the functional annotations from POKAY. This will create a new file which should replace the current file in assets/functional_annotation.

Acknowledgments

This workflow and scripts are written and conceptually designed by

Name	Affiliation
Zohaib Anwar; @anwarMZ	Centre for Infectious Disease Genomics and One Health, Simon Fraser University, Canada
Madeline Iseminger; @miseminger	Centre for Infectious Disease Genomics and One Health, Simon Fraser University, Canada
Anoosha Sehar; @Anoosha-Sehar	Centre for Infectious Disease Genomics and One Health, Simon Fraser University, Canada
Ivan Gill; @ivansg44	Centre for Infectious Disease Genomics and One Health, Simon Fraser University, Canada
William Hsiao; @wwhsiao	Centre for Infectious Disease Genomics and One Health, Simon Fraser University, Canada
Paul Gordon; @nodrogluap	CSM Center for Health Genomics and Informatics, University of Calgary, Canada
Gary Van Domselaar; @phac-nml	Public Health Agency of Canada

Many thanks to others who have helped out and contributed along the way too, including (but not limited to)*: Canadian COVID Genomics Network - VirusSeq, Data Analytics Working Group

Support

For further information or help, don't hesitate to get in touch at [email protected] or wwshiao

Citations

An extensive list of references for the tools used by the workflow can be found in the CITATIONS.md file.