Errors and warnings

When processing of a sequence fails for various reasons, not all output files will contain the corresponding entry (due to limitations of file formats):

Output file CLI arg Failed entries?
Nucleotide alignment --output-fasta no
Translations --output-translations no
Tree - Auspice v2 JSON --output-tree no
Tree - Newick --output-tree no
Genome annotation - GFF --output-annotation-gff no
Genome annotation - TBL --output-annotation-tbl no
Analysis results CSV --output-csv yes
Analysis results TSV --output-tsv yes
Analysis results NDJSON --output-ndjson yes
Analysis results JSON --output-json yes

You can find the reason for a particular failure by reading:

  • errors and warnings column in tabular outputs (TSV, CSV)

  • errors array in JSON output

  • error field in each entry in NDJSON output

Warnings

A warning does not stop the analysis of a sequence. You can find warnings in:

  • warnings column in tabular outputs (TSV, CSV), separated by ;

  • warnings array of each result in JSON and NDJSON outputs, with the fields cdsName and warning

  • tooltip of the sequence name in the results table of Nextclade Web

Nextclade reports these warnings:

  • CDS not translated: the CDS cannot be translated, for example because it is entirely deleted or not sequenced. The CDS is missing from the translation outputs.

  • Reverse complement: the sequence aligns only as its reverse complement. The cdsName of this warning is nuc.

  • Lost start codon: the CDS starts with ATG in the reference, but the first codon of the query is different, for example ATA, or -TG when a deletion removes part of the start codon. The CDS is still translated. A deletion over the start codon can be a real deletion, or an ambiguous deletion that the alignment placed into the CDS (see penaltyGapEnterCds in Sequence alignment). There is no warning when the first codon of the query contains N or ambiguous nucleotides, or is outside of the alignment range, because then the data do not show whether the start codon is present. There is also no warning for a CDS that continues the reading frame of another CDS, which is on the same strand and ends directly before it without a stop codon. This is the case for mature peptides of a polyprotein that the genome annotation lists as separate CDSs, for example HIV-1 p7 after p2, or dengue E after M. Translation runs into such a CDS, so its first codon is not a start codon, and Nextclade reports a change there only as an amino acid substitution. A mature peptide gets the warning when the CDS before it is not in the genome annotation, or is not selected with --cds-selection.