Basecalling workflow

Helper workflow for signal processing and primary data analysis of Oxford Nanopore Technologies' reads.

Introduction

In brief this workflow can be used to perform:

• Basecalling of a directory of pod5 or fast5 signal data
• Basecalling in Duplex mode
• Modified basecalling
• Basecalling in real time
• Output basecalled sequences in various formats: FASTQ, CRAM or Unaligned BAM
• If a reference is provided a sorted and indexed BAM or CRAM will be output for basecalling a directory of pod5 or fast5 signal data with dorado and aligning it with minimap2 to produce a sorted, indexed CRAM.

Compute requirements

Recommended requirements:

• CPUs = 64
• Memory = 256GB

Minimum requirements:

• CPUs = 8
• Memory = 64GB

Approximate run time: Variable depending on coverage, genome size, model of choice and GPU model.

ARM processor support: False

Install and run

These are instructions to install and run the workflow on command line. You can also access the workflow via the EPI2ME application.

The workflow uses Nextflow to manage compute and software resources, therefore nextflow will need to be installed before attempting to run the workflow.

The workflow can currently be run using either Docker or Singularity to provide isolation of the required software. Both methods are automated out-of-the-box provided either docker or singularity is installed. This is controlled by the -profile parameter as exemplified below.

It is not required to clone or download the git repository in order to run the workflow. More information on running EPI2ME workflows can be found on our website.

The following command can be used to obtain the workflow. This will pull the repository in to the assets folder of nextflow and provide a list of all parameters available for the workflow as well as an example command:

nextflow run epi2me-labs/wf-basecalling –-help

A demo dataset is provided for testing of the workflow. It can be downloaded using:

wget https://ont-exd-int-s3-euwst1-epi2me-labs.s3.amazonaws.com/wf-basecalling/wf-basecalling-demo.tar.gz
tar -xzvf wf-basecalling-demo.tar.gz

The workflow can be run with the demo data using:

nextflow run epi2me-labs/wf-basecalling \
    -profile singularity \
    --input wf-basecalling-demo/input \
    --ref wf-basecalling-demo/GCA_000001405.15_GRCh38_no_alt_analysis_set.fasta \
    --dorado_ext pod5 \
    --out_dir output \
    --basecaller_cfg [email protected] \
    --remora_cfg "[email protected]_5mCG@v2"

For further information about running a workflow on the cmd line see https://labs.epi2me.io/wfquickstart/

Related protocols

This workflow is designed to take input sequences that have been produced from Oxford Nanopore Technologies devices.

Find related protocols in the Nanopore community.

Input example

This workflow accepts a folder containing FAST5 or POD5 files as input. The folder may contain other folders of FAST5 or POD5 files and all files will be processed by the workflow.

Input parameters

Input Options

Nextflow parameter name	Type	Description	Help	Default
input	string	Directory containing FAST5 (or POD5) signal for basecalling.	This directory will be searched recursively. All FAST5 or POD5 files (depending on which extension you select in the Basecalling Options) in this directory or any subdirectory (no matter how deep) will be basecalled.
ref	string	Optional reference FASTA file to align basecalled reads to.	Without a reference, basecalls are output to unaligned CRAM. When using a reference, take care to retain this FASTA file as the output CRAM file cannot be read without the reference it was aligned to.

Output Options

Nextflow parameter name	Type	Description	Help	Default
out_dir	string	Directory for output of all files.		output
sample_name	string	Sample name to prefix file names of workflow outputs.		SAMPLE
fastq_only	boolean	Output unaligned FASTQ instead of unaligned CRAM.	FASTQ can only be output when a reference has not been provided. Aligned output will always be written to CRAM even if fastq_only is set.	False
output_bam	boolean	Output unaligned BAM instead of unaligned CRAM.	Some downstream applications do not yet support CRAM and will require a BAM file. Enabling this option will output BAM instead of CRAM. You should only use this option if you know that it is needed. Output files will be larger than the corresponding CRAM files that would have been written if this option was not enabled.	False

Basecalling options

Nextflow parameter name	Type	Description	Help	Default
basecaller_cfg	string	Name of the model to use for converting signal.	Required for basecalling. The model list only shows models that are compatible with this workflow.
duplex	boolean	Run the basecaller in duplex mode.	By default, the workflow conducts simplex basecalling. If you used a chemistry and flowcell combination that supported duplex reads, you should switch this option on. Currently, duplex basecalling is not compatible with modified basecalling. Additionally, duplex basecalling within this workflow is reliant on internal optimisations to organise input files for better duplex rates, which is not possible when using streaming basecalling; therefore duplex combined with the watch_path option could lead to lower duplex rates.	False
remora_cfg	string	Name of the model to use for calling modified bases.	Required for calling modified bases while basecalling. The model list only shows models that are compatible with this workflow.
dorado_ext	string	File extension for Dorado inputs.	Set this to fast5 if you have not converted your fast5 to pod5. It is recommended to convert existing fast5 files to pod5 for use with Dorado.	pod5

Advanced basecalling options

Nextflow parameter name	Type	Description	Help	Default
output_pod5	boolean	Save the converted POD5 when running in duplex with FAST5 inputs.	Dorado duplex only supports POD5 input. The workflow will automatically convert FAST5 input to POD5 when duplex calling. By default, converted POD5 are deleted to save disk space. Enabling this option will make the workflow output converted POD5 files to a subfolder within the output directory.	False
qscore_filter	number	Mean qscore by which to filter reads. Inclusive such that reads with score >= qscore_filter are kept.	The mean qscore of reads is calculated by dorado and rounded to an integer by dorado and stored as a tag in dorado's SAM output. The pipeline separates reads into pass and fail categories based on this SAM tag.	10
cuda_device	string	GPU device to use for basecalling [cuda:all].	For local execution this can be used to pin GPU tasks to one (or more) specific GPU devices. Use cuda:all to use all available GPU devices, or cuda:idx[,idx,...] where idx is an index number(s) of GPU device(s) to use.	cuda:all
basecaller_model_path	string	Override the named basecalling model with a custom basecalling model.	For typical use, users should set --basecaller_cfg which will use a named model from inside the container. Experimental or custom basecallers will not be available in the container and can be loaded from the host with --basecaller_model_path.
remora_model_path	string	Override the named remora model with a custom remora model.	For typical use, users should set --remora_cfg which will use a named model from inside the container. Experimental or custom models will not be available in the container and can be loaded from the host with --remora_model_path.
basecaller_basemod_threads	number	Number of threads to use for base modification calling.	You must set this to > 0 when using a modbase aware model. Modbase calling does not require much additional CPU and should be set carefully when using GPU servers with a small number of CPUs per GPU.	2
basecaller_args	string	Additional command line arguments to pass to the basecaller process.

Multiprocessing Options

Nextflow parameter name	Type	Description	Help	Default
ubam_map_threads	integer	Set max number of threads to use for aligning reads from uBAM (limited by config executor cpus)		8
ubam_sort_threads	integer	Set max number of threads to use for sorting and indexing aligned reads from uBAM (limited by config executor cpus)		3
ubam_bam2fq_threads	integer	Set max number of threads to use for uncompressing uBAM and generating FASTQ for alignment (limited by config executor cpus)		1
merge_threads	integer	Set max number of threads to use for merging BAM files (limited by config executor cpus)		4
stats_threads	integer	Set max number of threads to use for getting stats from output files. (limited by config executor cpus)		4

Real Time Analysis Options

Nextflow parameter name	Type	Description	Help	Default
watch_path	boolean	Enable to continuously watch the input directory for new input files. Reads will be analysed as they appear.	This option enables the use of Nextflow's directory watching feature to constantly monitor input directories for new files. As soon as files are written by an external process Nextflow will begin analysing these files. The workflow will accumulate data over time to produce an updating report. Real time analysis of duplex data may lead to lower duplex rates than what would have been obtained by running basecalling after sequencing.	False
read_limit	integer	Stop processing data when a particular number of reads have been analysed.	By default the workflow will run indefinitely when using the real time watch path option. This will set the upper bound on the number of reads that will be analysed before the workflow is automatically stopped and no more data is analysed.

Generic options

Nextflow parameter name	Type	Description	Help	Default
disable_ping	boolean	Disable workflow ping.		False

Outputs

Output files may be aggregated including information for all samples or provided per sample. Per-sample files will be prefixed with respective aliases and represented below as {{ alias }}.

Title	File path	Description	Per sample or aggregated
workflow report	wf-basecalling-report.html	Report summarising the work done by the basecalling workflow	per-sample
Simplex alignment file of passed reads	{{ alias }}.pass.simplex.{{ format }}	BAM or CRAM file of simplex reads for the sample that pass QC filtering.	per-sample
Duplex alignment file of passed reads	{{ alias }}.pass.duplex.{{ format }}	BAM or CRAM file of duplex reads for the sample that pass QC filtering. Created if duplex basecalling is requested.	per-sample
Simplex alignment file index of passed reads	{{ alias }}.pass.simplex.{{ format }}.{{ index_format }}	The index of the resulting BAM or CRAM file with the simplex reads that pass QC filtering.	per-sample
Duplex alignment file index of passed reads	{{ alias }}.pass.duplex.{{ format }}.{{ index_format }}	The index of the resulting BAM or CRAM file with the duplex reads that pass QC filtering. Created if duplex basecalling is requested.	per-sample
Simplex alignment file of failed reads	{{ alias }}.fail.simplex.{{ format }}	BAM or CRAM file of simplex reads for the sample that fail QC filtering.	per-sample
Duplex alignment file of failed reads	{{ alias }}.fail.duplex.{{ format }}	BAM or CRAM file of duplex reads for the sample that fail QC filtering. Created if duplex basecalling is requested.	per-sample
Simplex alignment file index of failed reads	{{ alias }}.fail.simplex.{{ format }}.{{ index_format }}	The index of the resulting BAM or CRAM file with the simplex reads that fail QC filtering.	per-sample
Duplex alignment file index of failed reads	{{ alias }}.fail.duplex.{{ format }}.{{ index_format }}	The index of the resulting BAM or CRAM file with the duplex reads that fail QC filtering. Created if duplex basecalling is requested.	per-sample

Pipeline overview

1. Prerequisites

The workflow uses Dorado for basecalling which includes the use of Remora for modified basecalling. Basecalling with Dorado requires an NVIDIA GPU with Pascal architecture or newer and at least 8 GB of vRAM.

Windows

Windows should not be considered as a supported operating systems for wf-basecalling as we do not directly support configuration of accelerated computing through WSL2 and Docker. Although we do not offer support, it is possible to set up Docker to use GPUs for most versions of Windows 11 and some versions of Windows 10 and we direct users to the CUDA on WSL User Guide. Users should take note of the support constraints section to ensure their environment is suitable before following the guidance. Do not install an NVIDIA driver into your WSL2 environment. Users are encouraged to download Dorado for Windows from the Dorado GitHub repository.

MacOS

MacOS should not be considered as a supported operating systems for wf-basecalling as we do not support accelerated computing through Docker on MacOS. On MacOS, GPU support through Docker remains in technical infancy. In addition, the containers we provide will not be able to leverage the M1 and M2 architecture and will not run as performantly as if Dorado had been run natively. Users are encouraged to download Dorado for MacOS directly from the Dorado GitHub repository.

Linux

When using Docker for accelerated computing on Linux, you will need the nvidia-container-toolkit installed. If you observe the error "could not select device driver with capabilities gpu", you should follow the instructions to install nvidia-container-toolkit here. You will need to follow the steps to:

• Setup the package repository and the GPG key (ignore the box about experimental releases)
• Update package listings
• Install nvidia-container-toolkit
• Configure the Docker daemon to recognize the NVIDIA Container Runtime
• Restart the Docker daemon to complete the installation after setting the default runtime

By default, workflows are configured to run GPU tasks in serial. That is, only one basecalling task will be run at a time. This is to prevent the GPU from running out of memory on local execution. When running workflows on a cluster, or in a cloud where GPU resources are isolated from one another, users should specify -profile discrete_gpus as part of the command invocation. This will allow for parallel execution of GPU tasks. You should ask your system administrator if you need to configure any additional options to leverage GPUs on your cluster. For example, you may need to provide a special string to the workflow's --cuda_device option to ensure tasks use the GPU assigned to them by the job scheduler.

2. Choosing a model

To select the relevant model see the dorado repository for a table of available models to select for --basecaller_cfg and --remora_cfg.

3. Aligning to a reference

The workflow can optionally perform the alignment of the basecalled data using minimap2 to a reference of choice, provided with the --ref option.

4. Duplex calling

wf-basecalling supports duplex calling, which is enabled with the --duplex option. If you used a chemistry and flowcell combination that supported duplex reads, you should switch this option on. The resulting BAM/CRAM will quality filtered and then automatically split in separate BAM/CRAM files for the simplex and duplex reads. Since dorado duplex requires the inputs to be in pod5 format, the workflow will perform the conversion automatically using pod5 convert fast5. These files are normally deleted upon completion of the analysis, but can optionally be saved by the user by providing the --output_pod5 option.

5. Real-time analysis

wf-basecalling can perform the basecalling as the pod5 files are generated. To enable this, provide the --watch_path option. The workflow will process the newly generated files as soon as they become available.

Troubleshooting

• Duplex mode cannot be used when calling modified bases. You must either run simplex basecalling with modified bases; or duplex calling without modified bases.
• Duplex mode with wf-basecalling is reliant on internal optimisations to organise input files for better duplex rates, which is not possible when using streaming basecalling; therefore duplex combined with the --watch_path option could lead to lower duplex rates than what would be achieved running the algorithm after sequencing is completed.

FAQ's

If your question is not answered here, please report any issues or suggestions on the github issues page or start a discussion on the community.

Related blog posts

• Importing third-party workflows into EPI2ME Labs

See the EPI2ME website for lots of other resources and blog posts.