Multiple values and length for the same circular RNA about sailfish-cir HOT 1 OPEN

I observed the same problem and it seems to be related to when genes on the opposite strand overlap the gene with the circRNA. In my cases the opposite genes are always fully contained inside an intron of the host gene.

From going through the code, it looks like first the exons of circRNA regions are extracted regardless of strand, and then exon set is reconstructed from the filtered circRNA exons. During the reconstruction the duplication occurs, if a circRNA region contains/overlaps a gene on the opposite strand

Currently, I consider blacklisting those "internal" genes and discard them from the input gtf.

from sailfish-cir.