Sequencing analysis pipeline (aqueduct) for validating plasmids and DNA assembly constructs, using long reads.
Set up credentials in the SCM configuration file, then pull the Nextflow image:
nextflow pull edinburgh-genome-foundry/Sequeduct
Pull the Docker image that contains the required software:
docker pull ghcr.io/edinburgh-genome-foundry/sequeduct:latest
Create a directory for your project and copy (or link) the FASTQ directories from your Nanopore run (e.g. into fastq
). Specify this together with a sample sheet in your commands:
# Preview
nextflow run edinburgh-genome-foundry/Sequeduct -r v0.2.0 -entry preview --fastq_dir='fastq' --reference_dir='genbank' \
--sample_sheet='sample_sheet.csv' \
-profile docker
# Analysis
nextflow run edinburgh-genome-foundry/Sequeduct -r v0.2.0 -entry analysis --fastq_dir='fastq' --reference_dir='genbank' \
--sample_sheet='sample_sheet.csv' \
--projectname='EGF project' \
-profile docker
# Review
nextflow run edinburgh-genome-foundry/Sequeduct -r v0.2.0 -entry review --reference_dir='genbank' \
--results_csv='results_finalised.csv' \
--projectname='EGF project review' \
--all_parts='parts_fasta/part_sequences.fasta' \
--assembly_plan='assembly_plan.csv' \
-profile docker
The above three commands each output a directory within results
. Similarly, Nextflow creates and uses a directory named work
, so ensure that your project directory doesn't have one. Specify revision of the project with -r
(a git branch or tag), and choose a configuration profile (with -profile
). Profiles are specified in the Nextflow config files.
Note that canu v2.2 requires minimum 100 reads, otherwise it returns an error. A fix has been posted, but it's not released yet.
For convenience, a script is included to collect plot files from the result directories (bin/collect_plots.py
).
git clone [email protected]:Edinburgh-Genome-Foundry/Ediacara.git
docker build . -f containers/Dockerfile --tag sequeduct_local
Copyright 2021 Edinburgh Genome Foundry
Sequeduct was written at the Edinburgh Genome Foundry by Peter Vegh.