Custom nf-core/nanoseq pipeline for Oxford Nanopore long-read sequencing, adapted for the UPWASZAK lab at EPFL: https://www.epfl.ch/labs/upwaszak

We provide four wrapper scripts for common use cases:

• Runs basecalling locally on a GPU-enabled machine (BMO)
• Transfers data and the pipeline to a remote server (TREX) via rsync
• SSHes into TREX, starts a tmux session, and runs the remaining steps (alignment, variant calling, annotation, etc.)

• Runs the full pipeline locally (basecalling + variant calling)

• Runs only the basecalling step locally

• Skips basecalling and runs all other steps locally (expects existing .fastq.gz input)

Each script has a configuration section you can customize. You must provide a valid input file (.pod5 or .fastq.gz), and tune parameters.

Example configuration section:

### === CONFIGURATION (edit these only) ===
REMOTE_HOST="upwaszaksrv1.epfl.ch"                
REMOTE_USER="${USER}"
SAMPLE_NAME="sampleName"  
REMOTE_DEST_DIR="${SAMPLE_NAME}"                   

INPUT_PATH="small_NA12878_DNA.pod5"  # .pod5 or .fastq.gz
GTF_PATH=""                                     
DORADO_MODEL="hac"                                 
DORADO_MODIFICATION="5mCG_5hmCG"
CALL_VARIANTS=true
VARIANT_CALLER="clair3"
CLAIR_MODEL="dorado_model"                         
STRUCTURAL_VARIANT_CALLER="longcalld"
PHASE_WHATSHAP=true
ANNOTATE_VCF=true
### =======================================

• Most parameters can be modified directly in the wrapper scripts (e.g., model type, input paths, tool settings).
• For more advanced configuration (e.g., max.cpus, max.memory, max.time, and default resources), check and edit the nextflow.config file.

• Main workflow: workflows/nanoseq.nf
• Organized into:
  • subworkflows/: groups of related steps
  • modules/: individual tool processes

To change behavior of a specific tool (e.g. Clair3):

1. Locate where it's called in nanoseq.nf
2. Follow to subworkflows/local/short_variant_calling.nf
3. Modify logic in modules/local/clair3.nf

Each module in the pipeline runs inside a container that includes all the necessary tools for that specific process (e.g., Clair3, LongCalld, etc.).

If you need a custom setup (e.g., additional tools, modified versions), you can assign a custom container directly inside the module by specifying:

container 'docker.io/ff1997/methylasso:latest'

To create your own container and push on Docker Hub:

1. Write a Dockerfile with the required tools.
2. Build and push it to Docker Hub:

docker buildx build --platform linux/amd64 -t ff1997/methylasso:latest --load .
docker push ff1997/methylasso:latest

All Dockerfiles used in this pipeline are stored under the /containers/ directory.

Francesco Feher
📧 [email protected]