Q20 rebasecalling

This pipeline is based on ONT instructions to rebasecall Q20+ samples.

It does simplex + 2 step duplex_tools + duplex call + MinIONQC.R

Dependencies

• Conda or Mamba
• Snakemake
• Guppy
• Duplex_tools
• Fastq_merge
• minion_qc

Config file

Example of config.yaml fill:

Flowcell_id: "FZZ0000"
RawData: "/data/path/Q20_RawData/FZZ0000"
modelFile: "dna_r10.4_e8.1_sup.cfg"
modelPath: "/software//guppy/ont-guppy_5.0.1X/data"
cuda: "cuda:0"
firstCall:
  gpu_runners: 20
secondCall:
  gpu_runners: 20
bin:
  guppy_bin: "/software/guppy/ont-guppy_5.0.1X/bin"
  guppy_duplex_bin: "/software/guppy/ont-guppy_0.0.0_duplexbeta/bin"
  duplex_tool: "/home/user/.local/bin/"
  minionqc: "/software/minion_qc/minion_qc-1.4.2/"
  fastq_merge_bin: "/software/fastq_merge/0.2.0/bin/"

It generates all output folders and outputs at current working directory. A best practice will be create a new folder called ${Flowcell_id}_Analysis and run the pipeline form it.

Run

snakemake -s /path/to/Snakefile --configfile config.yaml --use-conda  -j ${cores}

If you didn't installed mamba, you have to add --conda-frontend conda.

Extra

If you want to save space and time, you can run a first run, once, with this two paramaters extra:

 --conda-create-envs-only --conda-prefix /path/to/prefix

Then, you will specify in each run --conda-prefix and will use these environments.

If you copy the Snakefile you don't need to specify -s