ONT Basecalling / Demux Pipeline

Nextflow pipeline to perform basecalling and (optional) demultiplexing of ONT data, collect QC metrics, and generate a MultiQC report. Uses Dorado for basecalling and demultiplexing.

It's also possible to use the pipeline with data that has already been basecalled and/or demultiplexed (e.g., with MinKNOW). In this case, the pipeline will only perform QC and generate a MultiQC report. For this, set the basecalled_input parameter to true and provide the path to the basecalled data (usually fastq_pass).

Requirements

• Nextflow (>= 25.04)
• Apptainer / Singularity
• Dorado (0.9.0 tested). It can be used via container, or installed locally from https://github.com/nanoporetech/dorado.

Quick Start

1. Clone this repository: bash git clone https://github.com/catg-umag/ont-basecalling-demultiplexing
2. Demultiplexing setup (optional):
   • If demultiplexing is needed, create a samples.csv file containing at least the barcode and sample columns.
   • Ensure the barcode column includes the barcode identifier (e.g., barcode01), and the sample column lists the sample name, which will be used in reports and as the FASTQ filename.
3. Configure parameters:
   • Copy the example parameters file: bash cp params.example.yml my_params.yml
   • Modify myparams.yml according to your needs. Ensure that the `sampledataparameter points to yoursamples.csv` file if you are demultiplexing.
4. Run the pipeline: bash nextflow run ont-basecalling-demultiplexing/ -profile apptainer -params-file my_params.yml

Pipeline Parameters

| Parameter | Required | Default | Description | | -------------------------- | -------- | ---------------------------------- | --------------------------------------------------------------------------------------------------- | | experiment_name | No | - | Name of the experiment, used for reports (title and filename). | | data_dir | Yes | - | Path to the directory containing POD5 files. | | sample_data | No | - | Path to the CSV file containing the sample data (if not provided, will not perform demultiplexing). | | output_dir | No | results | Directory for saving results. | | basecalled_input | No | false | Specifies if the input data is already basecalled. | | fastq_output | No | true | Generates FASTQ files if true; otherwise, generates UBAM files. | | qscore_filter | No | 12 | Minimum QScore threshold for "pass" data, used in demultiplexing. | | dorado_basecalling_model | No | sup | Model used for basecalling. Check Dorado help for available options. | | dorado_basecalling_gpus | No | 1 | Number of GPUs to allocate for basecalling. | | dorado_demux_kit | No | EXP-NBD196 | Kit identifier used for demultiplexing. | | dorado_demux_both_ends | No | false | Demultiplexes using barcodes on both ends (5' and 3') if true. | | use_dorado_container | No | true | Uses Dorado via container if true; expects a local installation if false. | | qc_tools | No | ['fastqc', 'nanoq', 'toulligqc'] | Specifies which QC tools to run. Options: 'nanoq', 'nanoplot', 'fastqc', 'toulligqc', 'pycoqc'. |

Considerations

• The pipeline is compatible with SLURM clusters; use -profile slurm.
• GPU resources are required for basecalling. On SLURM, this pipeline will send jobs requesting GPUs with the --gres=gpu:X option.
• You can provide extra arguments to Dorado basecalling and demultiplexing using ext.args.