virpipa

VirPipa is a modular Nextflow DSL2 pipeline for HCV probe-capture assembly and reporting. It reproduces the bash-original results/ contract while adding a maintainable workflow graph, fixture-backed module tests, VADR annotation, and geno2pheno resistance annotation.

Virpipa hopperiensis

Overview

The pipeline takes paired-end FASTQ input, removes host contamination, selects the best HCV reference, builds a hybrid assembly, polishes it with a pilon loop, creates final consensus and CRAM outputs, annotates with VADR, and reports subtype and resistance calls.

Main outputs per sample include:

• polished consensus FASTA and FAI
• 0.15-iupac consensus FASTA
• best-reference FASTA, CRAM, and report branch
• filtered pilon-m* VCF set
• CRAM / CRAI files
• subtype BLAST outputs
• VADR GFF and BED
• coverage, report, and nucleotide-frequency TSVs
• geno2pheno resistance TSV / BED / by-drug TSV

Workflow

FASTQ
  -> hostile host removal
  -> subsample for best-reference selection
  -> map to all HCV references
  -> select best reference
  -> SPAdes assembly
  -> hybrid reference build
  -> pilon polishing loop
  -> pilon BAM -> majority-call replacement FASTA
  -> pilon VCF -> filtered pilon-m* VCFs
  -> 0.15-iupac consensus
  -> no-opt remap to 0.15-iupac
  -> final CRAM / report / BLAST branches
  -> VADR annotation
  -> geno2pheno resistance annotation
  -> finalize bash-style results/

Quick Start

Hopper

Push your current branch to the Hopper bare repo before starting a workflow there:

git push --set-upstream hopper refactor/modular-take3

Run the standard Hopper wrapper from the cluster login node:

ssh rs-fe1 'cd /fs1/jonas/src/virpipa && bash -l runme.sh resume'

For a test-module run on Hopper:

bash runme_test.sh resistance

Local Laptop

Use the skrotis environment for local development:

source ~/miniforge3/etc/profile.d/conda.sh
conda activate skrotis

Run a local module test:

nextflow run test_module.nf -profile local --module resistance

Run the full workflow locally with containers:

export SENTIEON_LICENSE=localhost:8990
NXF_OFFLINE=true nextflow run . \
  -profile local_containers \
  --input /tmp/virpipa_contract_local_samplesheet.csv \
  --outdir /tmp/virpipa_contract_results_local

Inputs

Samplesheet columns:

• clarity_sample_id or equivalent sample id column
• read1
• read2
• optional sample_name or lid
• optional run_name or sequencing_run

Example:

clarity_sample_id,read1,read2,sample_name,run_name
SAMPLE001,/path/to/R1.fastq.gz,/path/to/R2.fastq.gz,TEST001,test_run

Outputs

Sample outputs are published under:

• results/<run_name>/<sample_id>/

The default --publish_mode routine writes the flat Virtitta-compatible sample archive. Use --publish_mode debug to keep the same flat final files and additionally publish intermediate subdirectories such as bam/, vcf/, fastq/, spades/, mummer/, and pilon/.

If enabled with --pipeline_info, Nextflow run metadata is written under:

• results/<run_name>/pipeline_info/timeline.html
• results/<run_name>/pipeline_info/report.html
• results/<run_name>/pipeline_info/trace.txt

If --run_name is not provided, the built-in metadata path falls back to results/pipeline_info_<timestamp>/. DAG rendering is disabled by default; add --pipeline_info_dag to request dag.svg when Graphviz is available.

See docs/output.md for the current output contract and docs/pipeline_workflow.md for workflow details.

Resistance Rules

geno2pheno HCV rules are refreshed manually outside Hopper. The offline refresh entrypoint is:

python scripts/update_geno2pheno_rules.py --output-csv assets/hcv_geno2pheno_rules.csv

The pipeline currently consumes the committed CSV rules table by default via params.resistance_rules and the repo default is assets/hcv_geno2pheno_rules.csv.