Meta-issue - tidying up output GTFs

[SamBryce-Smith]

Collection of points to address for cleaning up GTFs:

• gene IDs are duplicated when have multiple transcripts sharing a boundary with novel last exons. should switch to a 'g1;g2 tx1_g1,tx2_g1;tx1_g2,tx2_g2' kind of approach