Including unannotated transcripts in ballgown diffexpression analysis

Question

[gabrielpenun]

I am interested in some potential novel transcripts in an RNA-Seq data set, so I am producing a transcriptome with stringtie that does not rely on the gtf. I generate stringtie output that contains transcripts without a reference_id and was able to identify an unannotated transcript of interest, but cannot find this transcript in the ballgown input tables. I read over the function of the -G and -B flags of stringtie and can't see why the unannotated transcripts would not be included.