-[nf-core/viralrecon] Pipeline completed with errors- Error executing process > 'NFCORE_VIRALRECON:NANOPORE:ARTIC_MINION (SRX12373518)' Caused by: Process `NFCORE_VIRALRECON:NANOPORE:ARTIC_MINION (SRX12373518)` terminated with an error exit status (20) about viralrecon HOT 2 OPEN

I think this is a similar issue I outlined in more detail here artic-network/fieldbioinformatics#125 with the artic minion step.

You could examine your vcf files and see if it seems similar.

Essentially it seems there are overlapping variants in the fail.vcf and pass.vcf which causes preconsensus fasta file to be masked and thus bcftools complains when applying the "pass" variant as it expects the reference but instead finds the masked "N".

Typically I have seen this happen in indels when an amplicon drops out - when medaka variant is run on the pool containing the dropped amplicon it prodcues variants that should be filtered out - but dont seem to be.

I was wondering if it was something odd/bug happening with the longshot annotation and dropped amplicons (as it doesnt happen when medaka annotate is used via the --no-longshot flag in minion) causing the artifact variants to be kept from the dropped amplicon. Unforunately I havent had to time to look further.

One way around the issue is to run minion manually using the --no-longshot flag or use bcftools norm with the --check-ref x flag to exclude those sites.

from viralrecon.

@dejenie21 I think that @Sam-Sims might be onto something.

BTW, the second script you run seems to have issue accessing file on github given the errors

from viralrecon.