Tag "transcript_id" not found in attributes when in index with vg

[ld9866]

Dear developer:

Since it takes a very long time for vg to compare transcriptome data to pangenome data, about 5 days for each sample of 32 cores, which may be caused by too many variations in it, I want to extract a gene region for analysis, but we found that GTF could not be input after extracting it separately.

I read all the previous posts and I still don't think I know what to do. Can you help me.

Best regards,
Dong

The code:
vg autoindex --threads 32 --workflow mpmap --workflow rpvg --prefix vg_rpvg --ref-fasta test.fa --vcf test.vcf.gz --tx-gff test.gtf

The error:
[IndexRegistry]: Checking for phasing in VCF(s).
[IndexRegistry]: Chunking inputs for parallelism.
[IndexRegistry]: Constructing spliced VG graph from FASTA and VCF input.
ERROR: Tag "transcript_id" not found in attributes (line 3).

The data:

test.tar.gz

[jeizenga]

The 9th column in a GTF file is a semi-structured list of named fields and values, and typically, one of these fields indicates a unique identifier for the transcript, such as an accession number/ID. In GENCODE, that field is called "transcript_id", so that is the default in vg autoindex and vg rna. However, if you are using annotations from a different source, they often have a different name for this field. You'll probably be able to figure out which field is the identifier if you look at line 3 of the GTF (as the error indicates) and then you can provide the name of that field to vg autoindex with the --gff-tx-tag option.