Nanoporcini

Nanoporcini is an experimental pipeline for long-read metabarcoding of fungi with Oxford Nanopore Technologies sequencing.

Features

Nanoporcini is written in the Nextflow workflow language and features:

• Container support for Docker or Singularity.
• Configurable quality filtering with cutadapt and chopper.
• Full ITS region extraction with itsxpress.
• Chimera detection with VSEARCH
• A choice of clustering approaches:
  • VSEARCH
  • Custom implementation of NanoCLUST
• Taxonomic classifications using dnabarcoder

Requirements

• Nextflow v23.10.0 or later
• A container platform. Either:
  • Docker
  • or Singularity
• A network connection to download dependencies
• A reference database containing ITS sequences
  • UNITE+INSD 2024

Quick start

Run with minimal required parameters

nextflow run https://github.com/aringeri/nanoporcini \
  --input "path/to/data/sample*.fastq.gz" \
  --outdir "nanoporcini_output/" \
  --primers.fwd "AACTTAAAGGAATTGACGGAAG" \ 
  --primers.rev_rc "GGTAAGCAGAACTGGCG" \
  --chimera_filtering.ref_db "path/to/unite/db.fasta" \
  --taxonomic_assignment.dnabarcoder.ref_db "path/to/unite/db.fasta" \
  --taxonomic_assignment.dnabarcoder.ref_classifications "path/to/unite/db.classification" \
  --taxonomic_assignment.dnabarcoder.cutoffs "path/to/unite/db/cutoffs.best.json" 

This example uses NS5 forward primer and LR6 reverse primer (see https://unite.ut.ee/primers.php).

Test run with example data

A small example can be run with two samples and 20 reads each. (see tests/test_data/example/fq). This takes ~25 minutes to run on a 4 core macbook pro with 16GB RAM.

First clone the repository and move into the nanoporcini directory:

git clone https://github.com/aringeri/nanoporcini
cd nanoporcini 

Download the UNITE reference database using the command below or do it through your browser at https://zenodo.org/records/12580255. This may take a while:

mkdir -p data/db/unite2024/
curl https://zenodo.org/api/records/12580255/files-archive -o data/db/unite2024/unite2024.zip
unzip data/db/unite2024/unite2024.zip -d data/db/unite2024

mkdir -p data/db/dnabarcoder/
curl https://raw.githubusercontent.com/vuthuyduong/dnabarcoder/refs/heads/master/data/UNITE_2024_cutoffs/unite2024ITS.unique.cutoffs.best.json \
  -o data/db/dnabarcoder/unite2024ITS.unique.cutoffs.best.json

Run the examples:

nextflow run main.nf \
  -params-file "tests/test_data/example/example-params.yml" \
  --outdir "nanoporcini_output/" \
  -c conf/envs/local.config

If you are running from a HPC environment change the -c conf/envs/local.config line to -c conf/envs/cluster.config. This will allow the pipeline to use more resources and execute tasks using the Slurm workload manager.

Configuration

Many of the pipeline parameters can be configured to suit your needs. I recommend using creating a yaml file to pass parameters to the pipeline. See conf/params.yml for an example configuration file that can be given to the run command with the -params-file option:

nextflow run https://github.com/aringeri/nanoporcini \
  --input "path/to/data/sample*.fastq.gz" \
  -params-file conf/params.yml \
  --outdir "nanoporcini_output/"

Inputs

Samples

This pipeline is expecting ONT sequences to be basecalled and demultiplexed already. I recommend Dorado for these steps.

• input - Required
  • type - string
  • A set of fastq files. One file per sample. File names will be used to identify samples. Uses glob syntax (*) to select multiple files.
  • ex) "path/to/input/data/*.fastq.gz"

Primer sequences

• primers.fwd
  • type - string
  • The forward primer sequence in 5' to 3' orientation.
  • ex) "AACTTAAAGGAATTGACGGAAG" for NS5 primer (see https://unite.ut.ee/primers.php)
• primers.rev_rc
  • type - string
  • The reverse primer sequence (which has been reverse complemented).
  • ex) "GGTAAGCAGAACTGGCG" for LR6 primer (see https://unite.ut.ee/primers.php)

Outputs

Specify the output directory with:

• outdir
• qc_plot_sample_level

Quality Filtering

• FULL_ITS:
  • minQualityPhred: 20
  • minLength: 300
  • maxLength: 6000
• chimera_filtering:
  • ref_db

Clustering

• methods
  • VSEARCH
    • min cluster size
  • NanoCLUST
    • min cluster sizes

Taxonomic Assignments

• dnabarcoder
  • ref_db
  • ref_classifications
  • cutoffs

Customising Resource Use (CPU/RAM)

For controlling the number of threads or RAM usage of various tasks see the configuration files in conf/envs. These will allow you to use the available resources on your system and improve the pipeline runtime. Pass this to the run command with the -c option:

nextflow run https://github.com/aringeri/nanoporcini \
  --input "path/to/data/sample*.fastq.gz" \
  -c conf/envs/local.config \
  --outdir "nanoporcini_output/"

Further reading

This workflow was developed as part of my Masters degree. My thesis can be read online (https://aringeri.github.io/long-read-ITS-metabarcoding-thesis/), and it outlines the details of the pipeline, reasoning behind design choices and the validation approaches taken.