transcript loss with gff2gbSmallDNA.pl

[lazlostrange]

There appears to be an issue with the script gff2gbSmallDNA.pl. When applying gff2gbSmallDNA.pl to GeneMarkS-T predictions made on transcripts assembled by StringTie2, fewer transcripts are output than are input. This issue extends to converting the transcripts back to GFF-format with gbSmallDNA2gff.pl. In this conversion, an additional number of transcripts are lost.

You can find example data in /nas-hs/projs/braker/longreads/gff2gb_sandbox. How the data was generated is explained in make.doc.