NC_045512 shows up in MAT when extracting from a WNV auspice tree

[jcw349]

Hi,

I am trying to convert a West Nile virus auspice tree to a MAT. The matUtils extract keeps including "NC_045512" in the output file, but it's not in my input auspice file.

matUtils extract -i auspice/WNV-global.json -o usher/WNV-global.pb

I tried specifying the reference files and metadata as well, but it's still creating the same MAT.
matUtils extract -i auspice/WNV-global.json -g config/reference.gtf -f config/reference.fasta -o usher/WNV-global.pb

reference: NC_009942

VCF from the output MAT:
matUtils extract -i usher/WNV-global.pb -g config/reference.gtf -f config/reference.fasta -v usher/mutations.vcf

First 5 rows and 12 columns of the vcf:

##fileformat=VCFv4.2
#CHROM  POS     ID      REF     ALT     QUAL    FILTER  INFO    FORMAT  AY765264        KJ831223        FJ159130
NC_045512       1       A1C,A1G,A1T     A       C,G,T   .       .       AC=16,67,3;AN=4560      GT      0       0       0
NC_045512       2       G2A,G2C,G2T     G       A,C,T   .       .       AC=12,4,20;AN=4560      GT      0       0       0
NC_045512       3       T3A,T3C,T3G     T       A,C,G   .       .       AC=89,25,4;AN=4560      GT      1       0       0

Not sure what I'm doing wrong.

Thank you,
Jade W

[AngieHinrichs]

Sorry about that @jcw349! When writing VCF, "NC_045512" is hardcoded! We should be able to do better than that.

Hopefully something like this will work for you in the meantime?:

sed -e 's/^NC_045512/NC_009942/;' usher/mutations.vcf > usher/mutations.renamed.vcf

[jcw349]

Sorry about that @jcw349! When writing VCF, "NC_045512" is hardcoded! We should be able to do better than that.

Hopefully something like this will work for you in the meantime?:

sed -e 's/^NC_045512/NC_009942/;' usher/mutations.vcf > usher/mutations.renamed.vcf

No worries!! Thank you for looking into this and sharing a solution to fix the VCF.

The matUtil extract -i <json_file> -o <mat.pb> also seems to be labeling non-covid trees the same reference, NC_045512. It was in the output MAT file too. Not sure if that'll have a big impact on doing other things with the file?

For now I remade the MAT.pb using UShER, which did end up using NC_009942, but it didn't keep all of the same things I had initially put into the nextstrain tree like filters and colors, etc.

[AngieHinrichs]

Yes, the MAT protobuf contains only the mutation annotated tree, not the other many things that can be layered onto Nextstrain's Auspice JSON format. The matUtils extract options for adding in a reference, metadata etc. are only used when generating JSON output AFAIK.

[jcw349]

Sorry, what I mean is, when I tried to use matUtil extract to convert the Auspice json to MAT protobuf, for some reason the MAT.pb ended up having the NC_045512 in it too, even though that's not in the input json.

[AngieHinrichs]

Yes, NC_045512 is hardcoded when importing JSON, sorry. Does auspice/WNV-global.json contain "NC_009942" anywhere in it? (If you're able to share auspice/WNV-global.json privately then I can take a look at how matUtils might figure out what the reference name should be.)