Citation:
poreCov - an easy to use, fast, and robust workflow for SARS-CoV-2 genome reconstruction via nanopore sequencing
Christian Brandt, Sebastian Krautwurst, Riccardo Spott, Mara Lohde, Mateusz Jundzill, Mike Marquet, Martin Hölzer
https://www.frontiersin.org/articles/10.3389/fgene.2021.711437/full
- poreCov is a SARS-CoV-2 analysis workflow for nanopore data (via the ARTIC protocol) or SARS-CoV-2 genomes (fasta)
- the workflow is pre-configured to simplify data analysis:
- poreCov | SARS-CoV-2 Workflow for nanopore sequencing data
- Table of Contents
- 1. Quick Setup (Ubuntu)
- 2. Run poreCov
- 3. Quality Metrics (default)
- 4. Workflow
- 5. Literature / References to cite
- 6. Troubleshooting
- 7. Time to results
- 8. Credits
- poreCov needs Nextflow and java run time (default-jre)
- install java run time via:
sudo apt install -y default-jre - install Nextflow via:
curl -s https://get.nextflow.io | bash && sudo mv nextflow /bin && sudo chmod 770 /bin/nextflow
- install java run time via:
- installation here (recommended), alternatively via:
sudo apt install -y docker - add Docker to the user:
sudo usermod -a -G docker $USER
- Singularity installation here
- if you can't use Docker
- Conda installation here
- install Nextflow and Singularity via conda (not cluster compatible) - and use the
singularityprofile
- only important if you want to do basecalling via GPU with the workflow:
- local guppy installation (see oxford nanopore installation guide)
- or: install nvidia Docker tool kit
- or: Singularity (with --nv support)
- validate your installation via test data:
# for a Docker installation
nextflow run replikation/poreCov -profile test_fastq,local,docker -r 0.11.0
# or for Singularity or conda installation
nextflow run replikation/poreCov -profile test_fastq,local,singularity -r 0.11.0- poreCov with basecalling and Docker
--updatetryies to force the most recent lineage release version (optional)-r 0.11.0specifies the workflow release from here
nextflow run replikation/poreCov --fast5 fast5/ -r 0.11.0 \
--cores 6 -profile local,docker --update - poreCov with a basecalled fastq directory
nextflow run replikation/poreCov --fastq_pass 'fastq_pass/' -r 0.11.0 \
--cores 32 -profile local,docker- poreCov with basecalling and renaming of barcodes based on
sample_names.csv
# rename barcodes automatically by providing an input file, also using another primer scheme
nextflow run replikation/poreCov --fast5 fast5_dir/ --samples sample_names.csv \
--primerV V1200 --output results -profile local,docker- see also
nextflow run replikation/poreCov --help -r 0.11.0
- poreCov supports version control via
-rthis way, you can run everything reproducible (e.g.-r 0.11.0) - poreCov releases are listed here
- add
-r <version>to a poreCoV run to activate this - run
nextflow pull replikation/poreCovto install updates- if you have issues during update try
rm -rf ~/.nextflowand thennextflow pull replikation/poreCov - this removes old files and downloads everything new
- if you have issues during update try
- these are the flags to get "data" into the workflow
--fast5 fast5_dir/for fast5 directory input--fastq_pass fastq_dir/directory with basecalled data (contains "barcode01" etc. directories)--fastq "sample*.fastq.gz"alternative fastq input (one sample per file)--fasta "*genomes.fasta"SARS-CoV-2 genomes as fasta (.gz allowed)
- barcodes can be automatically renamed via
--samples sample_names.csv - required columns:
_id= sample nameStatus= barcode number which should be renamed
- optional column:
Description= description column to be included in the output report and tables
Example comma separated file (don't replace the header):
_id,Status,Description
Sample_2021,barcode01,good
2ndSample,BC02,bad
- lineage determinations are quickly changing in response to the pandemic
- to avoid using out of date lineage schemes, a
--updateflag can be added to each poreCov run to get the most recent version-controlled pangolin container - we are currently building two times every week version-controlled pangolin container automatically, see here
- it is also possible to use instead of
--updatethis flag:--pangolindocker 'nanozoo/pangolin:3.1.1--2021-06-14' - this way you can use other container or version, but beware some containers might not be compatible with poreCov
- it is also possible to use instead of
- Regions with coverage of 20 or less are masked ("N")
- Genome quality is compared to NC_045512.2
- Genome quality assessment is based on RKIBioinformaticsPipelines/president
- also prepares csv and fasta for upload via DESH portal
- Genome quality assessment is based on RKIBioinformaticsPipelines/president
- Pangolin lineages are determined
- nextstrain clades are determined including mutation infos
- reads are classified to human and SARS-CoV-2 to check for possible contamination and sample prep issues
- poreCov was coded with "easy to use" in mind, while staying flexible
- therefore we provide a few input types which adjusts the workflow automatically (see image below)
- fast5 raw data, fastq files (one sample per file), fastq_pass (the basecalling output) or fasta (supports multifastas)
- primer schemes for ARTIC can be V1, V2, V3(default), V4, V4.1 or V1200 (the 1200bp amplicon ones)
For citing etc. check out these programs used for poreCov:
- Collection of some helpful infos
- Singularity needs additional option flags to run like
--usernsSolution on how to pass Singularity commands to poreCov
Table 1: Execution speed of poreCov on different Ubuntu 20 Systems using a single sample file with 167,929 reads. Command used: nextflow run replikation/poreCov -r 0.9.4 -profile test_fastq,local,docker.
| Hardware | First time with download (DB+container)¹ | Default settings |
|---|---|---|
| 2 CPUs 4 GB RAM | 1h 2min | 32 min 30s ² |
| 2 CPUs 8 GB RAM | 46 min | 21m 20s |
| 4 CPUs 16 GB RAM | 40 min | 12m 48s |
| 8 CPUs 32 GB RAM | 35 min | 11m 39s |
| 16 CPUs 64 GB RAM | 30 min | 9m 39s |
¹ time depends mostly on available internet speed
² was not able to execute read classification due to limited hardware, but generated and classified SARS-CoV-2 genomes
Table 2: Execution speed of poreCov on different Ubuntu 20 Systems using 24 fastq samples. Command used: nextflow run replikation/poreCov -r 0.9.4 --fastq "*.fastq.gz" --primerV V1200 --samples samplenames.csv -profile local,docker. Time meassured by the start of the workflow.
| Hardware | Default settings |
|---|---|
| 2 CPUs 4 GB RAM | 13h 33m ¹ |
| 2 CPUs 8 GB RAM | 7h 56m ¹ |
| 4 CPUs 16 GB RAM | 4h 10 min |
| 8 CPUs 32 GB RAM | 2h 15 min |
| 16 CPUs 64 GB RAM | 1h 25 min |
¹ was not able to execute read classification due to limited hardware, but generated and classified SARS-CoV-2 genomes
The key steps of poreCov are carried out using the ARTIC Network field bioinformatics pipeline. Kudos to all amazing developers for your incredible efforts during this pandemic! Many thanks to all others who have helped out and contributed to poreCov as well.