v1.0.0

Added

• RNA004 models
• R941 v3.3 5mCG 5hmCG models
• Duplex calling with option --duplex
  • Note that duplex calling is not optimised for streaming basecalling with --watch_path and may lead to lower duplex yield.
  • Duplex basecalling is currently not compatible with modified basecalling.

Changed

• Updated Dorado to v0.3.2 (see https://github.com/nanoporetech/dorado/releases/tag/v0.3.2)
• Pascal architecture GPUs are now supported
• Bumped minimum required Nextflow version to 23.04.2
• Users no longer need to provide --basecaller_cfg custom and/or --remora_cfg custom to override models with --basecaller_model_path and/or --remora_model_path respectively.

Fixed

• bamstats process very slow when output_bam has been selected

v0.7.2

Added

• v4.2 5mC and 6mA modification models

Changed

• Updated Dorado to v0.3.1
• GPU tasks are limited to run in serial by default to avoid memory errors
  • Users in cluster and cloud environments where GPU devices are scheduled must use -profile discrete_gpus to parallelise GPU work
  • A warning will be printed if the workflow detects it is running non-local execution but the discrete_gpus profile is not enabled
  • Additional guidance on GPU support is provided in our Quickstart
• Bumped minimum required Nextflow version to 22.10.8

v0.7.1

Fixed

• Command not found on cram_cache step
• Typo in report that refers to the workflow as "wf-basecalling-report"

v0.7.0

Changed

• Updated Dorado to v0.3.0
• BAM may be output instead of CRAM by providing --output_bam
• --help message will list basecalling and modbasecalling models available for use with the workflow

Added

• v4.2.0 models, which must be used for sequencing runs performed at new 5 kHz sampling rate
• v4.1.0 models replace v4.0.0 models and must be used for sequencing runs performed at 4 kHz sampling rate

Removed

• v4.0.0 models

Fixed

• Custom models were previously rejected by the workflow as basecaller_cfg and remora_cfg are validated against a list of basecalling models installed in the Dorado container.
  • Users should now provide --basecaller_cfg custom and/or --remora_cfg custom to override models with --basecaller_model_path and/or --remora_model_path respectively.
  • Providing --basecaller_cfg custom or --remora_cfg custom without the corresponding --basecaller_model_path or --remora_model_path will result in an error.

v0.6.0

Added

• Ability to watch the input path and process files as they become available in real time.

v0.5.2

Added

• Configuration for running demo data in AWS

v0.5.1

Fixed

• Missing models from list of valid models
• "[email protected]_5mCG@v0" is now correctly referred to as "[email protected]_5mCG@v0", to match the simplex model version
• "[email protected]_5mCG@v0" is now correctly referred to as "[email protected]_5mCG@v0", to match the simplex model version

v0.5.0

Changed

• Updated Dorado to v0.2.4
• Updated to Oxford Nanopore Technologies PLC. Public License

Fixed

• Dorado image correctly ships with CUDA runtime library

v0.4.1

Fixed

• Input ref channel depleted after first alignment

v0.4.0

Changed

• Reference is no longer required for basecalling
  • CRAM files with no alignments will be generated if --ref is not provided
  • FASTQ may be output instead of CRAM by providing --fastq_only
• PG line for converting Dorado SAM output to uBAM is no longer written to output header
• Work directory is automatically cleaned up on successful completion to remove large intermediate files
  • Override this by including cleanup = false in a custom Nextflow configuration file
• Number of threads for merging is now configurable for advanced users