genomebam processing does not check for missing transcript_id

[SolKatzman]

kallisto v0.48.0

When run with the --genomebam option, kallisto requires a gtf file to map the coordinates of the exons for each transcript that is one of the targets in the fasta file used to generate the kallisto target index.

I ran the program with a gtf that did not contain transcript records, on the assumption that the program would find the transcript_id in the attributes of each exon record in the gtf. The genomebam output of this run had all of the reads unmapped, with the chromosome indicated as "*".

But even with transcript records, it is possible due to typos or other errors that there might be missing transcript_id for a given fasta ID. Rather than silently producing erroneous output, it would be preferable for kallisto to immediately crash with an error message indicating "in gtf (filename), no transcript_id found corresponding to fasta id XXX in the target index"

Thank you for your attention.

Sol Katzman UCSC Genomics Institute