vg rna error

[GSgen]

Hello,

I have made pggb graph for a gene and I am trying to run vg rna using:

vg convert -g smooth.gfa -p > example.pg
vg index example.pg -x example.xg
vg snarls example.xg > example.snarls
vg deconstruct example.xg -r example.snarls -d 2 -e -a -p REF Sample1 Sampl2 Sample3 > example.vcf
vg rna -p -t 16 -n example.gff example.pg > example_spliced.pg

However, it crashes:

vg rna -p -t 16 -n example.gff example.pg > example_spliced.pg
[vg rna] Parsing graph file ...
[vg rna] Graph parsed in 0.00121847 seconds, 0.00424194 GB
[vg rna] Adding transcript splice-junctions and exon boundaries to graph ...
vg: src/transcriptome.cpp:640: void vg::Transcriptome::add_exon(vg::Transcript*, const std::pair<int, int>&) const: Assertion `exon_pos.second < transcript->chrom_length' failed.
ERROR: Signal 6 occurred. VG has crashed. Visit https://github.com/vgteam/vg/issues/new/choose to report a bug.
Stack trace path: /tmp/vg_crash_yzTBqF/stacktrace.txt
Please include the stack trace file in your bug report!

My goal is to make a pan-transcriptome after making pan-genome using vg rna. Could you please help to fix it? Thank you

[jonassibbesen]

Hi, based on the assert it looks like one of the exons in the example.gff transcript annotation has an end coordinate that is larger than the length of the corresponding reference path in the graph. You can check whether this is the issue by first running vg paths -E -v example.pg. This will give you the length of all paths in the graph, which you can then compare to the exon coordinates in the annotation.