We converted the tutorial into a very simple Nextflow pipeline. The code is available from Github

What do you need to run

The pipeline will need

• your assembly (FASTA),
• a set of paired end short reads,
• and the databases (for Genomad and CheckV), that now you know you can download with votuderep.

The reads are to be supplied as a mapping file that you can create with votuderep tabulate.

votuderep tabulate ~/virome/reads -o ~/reads.csv

Executing the pipeline

One of the cool things about Nextflow is that you don’t need to download the pipeline from github, and you don’t need to install the dependencies: they will be fetched using your favourite source (in our case, Docker). We could use Nextflow also to download the databases…

nextflow run quadram-institute-bioscience/nf-ebame-virome -profile docker \
  --assembly  ~/virome/human_gut_assembly.fa.gz \
  --reads     ~/reads.csv \
  --genomad_db  ~/db/genomad_db/ \
  --checkv_db  ~/db/checkv-db-v1.5/ \
  --outdir    ~/nf-ebame-virome-out 

After you hit enter, Nextflow will download the pipeline and start orchestrating the tasks

Output

The output folder will contain a subdirectory for each task (it’s still a bit messy):

output-directory
├── genomad
│   └── genomad-out
│       ├── input_assembly_aggregated_classification
│       ├── input_assembly_annotate
│       │   └── input_assembly_mmseqs2
│       │       ├── ...
│       ├── input_assembly_find_proviruses
│       │   └── input_assembly_provirus_mmseqs2
│       │       ├── ...
│       ├── input_assembly_marker_classification
│       ├── input_assembly_nn_classification
│       │   ├── ...
│       └── input_assembly_summary
├── dereplicated-votus           (dereplicated vOTU FASTA)
├── checkv
│   └── checkv-out
│       └── tmp...
├── alignments                   (your BAM files)
└── pipeline_info