Nanowgs

This repository contains the Nextflow pipeline used to analyse Nanopore whole-genome sequences in the context of ASAP. Current version as of 2023-08-08

Nanowgs on wice

To run nanowgs on VSC HPC wICE cluster.

• First, clone the repository

git clone https://github.com/AlexanRNA/nanowgs.git

• Next, you need to update the config file. Update the following parameters. They can also be overwritten by command line input parameters

genomeref 
sampleid
guppy_config OR dorado_config
mod_bases
clair3_config
shasta_config
shasta_minreadlength (6000 recommended)
outdir
karyotype 

• Update working directory. Preferably, you use your own scratch

workDir = '/scratch/leuven/path/to/scratch/nextflow'

• In singularity profile part of config, you may want to update the location of the singularity images. However, the current location should work for you

singularity.cacheDir = '/path/to/images'

• Once your config file is updated, you can proceed with actually running the pipeline

Open a new tmux window on one of the reserved Genius nodes (r23i27n14 r23i27n22 r23i27n23 r23i27n24), or any CPU wICE node. Login node is not recommended, since this pipeline runs for a while and can be thus killed at some point.

tmux new-session -A -s nanowgs

Load nextflow (version >=19.10.0) and export environmental variable. You can also save the environmental variable in your .bashrc file

ml Nextflow
export SLURM_CLUSTERS=wice

Now, you can run nanowgs nextflow pipeline.

Guppy + following analysis ( pre-Feb 2023 )

To run guppy basecalling and the following processing, use the following command. I recommend starting the pipeline in your output directory, so you have report and timeline, as well as all nextflow logs at the same location.

/path/to/repo/nanowgs/main.nf \
-profile singularity,slurm \
-entry slurm \
--karyotype xx \
--outdir /path/to/output_dir -with-report -with-timeline 

Dorado basecalling ( August 2023 )

To run new ONT basecaller, use the following command. It will first convert fast5 files to pod5 and then proceed to basecall using ONT Dorado (v0.3.2). Make sure you specify the model and modified bases required.

If no output directory is specified, the output will be located wherever the nextflow pipeline was started. You provide --fast5 in case your data is in fast5 format and not in pod5. In case of pod5, no parameter/flag needs to be provided.

You can also use rerio modified basecalling models. In that case, you need to update rerio_config with the native basecalling model and provide base modification you wish to call in mod_bases.

/path/to/repo/nanowgs/main.nf \
-profile singularity,slurm \
-entry dorado_call \
--dorado_config "[email protected]" \
--mod_bases "5mCG_5hmCG" \
--sampleid ASA_143B \
--ont_base_dir /path/to/fast5 \
--fast5
-with-report -with-timeline 

Following analysis (August 2023, still in development)

This is the command that can be used to run subsequent analysis. However, bear in mind, this part of pipeline is still in development and was not fully tested. The example below is specific for LSK114 chemistry basecalled with [email protected] basecalling model. In case your data also hase modified bases information, do include --mod_bases 5mC (5mC is just exmaple) to your command to get some statistics and visualisations around TSS/CTCF.

/path/to/repo/nanowgs/main.nf \
-profile singularity,slurm \
-entry slurm_dorado \
--ubam /location/of/your/file.bam \
--sampleid ASA_145B \
--outdir /path/to/output_dir \
--clair3_config "/opt/bin/rerio/clair3_models/r1041_e82_400bps_sup_v400" \
--shasta_config "Nanopore-R10-Fast-Nov2022" \
--karyotype xy \
-with-timeline -with-report

TODO

Todo

• make the bam alignment more effective (too time consuming atm)
• update dorado version + script
• investigate CRAM saving

In Progress

• modified basecalling QC/checking
• update sniffles version + size filtering to 50
• update clair3 version + size filtering to 50
• update longphase version

Done ✓

Cite this work