H3N2 NA aa length is wrong about seasonal-flu HOT 6 CLOSED

For now, I've manually edited the genome annotations in the current seasonal flu builds so you can use the color-by option.
Once #99 is merged, future builds should be fixed as well.

from seasonal-flu.

Out of interest is this an auspice setting or an issue with augur output?

from seasonal-flu.

Hi @ammaraziz, thank you for reporting this bug!

The root of the problem seems to be the CDS coordinates in our workflow's H3N2 NA reference. The current coordinates are 4..1410, but the full coordinates should be 4..1413 based on the record in GenBank.

This reference is passed to the augur translate command via the --reference-sequence option in the workflow. The features parsed from the reference are used to generate the genome annotations which are ultimately exported by augur export v2 as the meta.genome_annotations in the main Auspice JSON file.

Within Auspice, the range displayed is based on the gene length calculated from the meta.genomeAnnotations. With the current CDS coordinates, this is calculated as Math.floor((1410 - 4) / 3), which is 468.

from seasonal-flu.

Notably, even in the copy we have with the bad coordinates, the translation field for that CDS is the correct length (469).

This also raises the question of if any of the other reference files are subtly wrong?

from seasonal-flu.

Thank you for the quick fix @joverlee521. I was trying to track where the issue was coming from, so thank you for explaining.

@tsibley Good point. I visually inspected (using UGENE software) HA/NA segments for the 4 subtypes. From how the range is calculated, the stop codon needs to be included.

The other NA gb files are correct (except the aforementioned issue with H3)
H3HA short by 3 for HA2

Thank you for fixing this issue!

from seasonal-flu.

Fixed by #99

from seasonal-flu.