RABT assembly output

[Trudy]

Hello!

I've a question about the output of cufflinks with RABT assembly.
In the isoforms.fpkm_tracking file I have 3 different situations:

1) ENST00000492842 - - ENSG00000240361 if the software identified transcripts that are included in the gtf reference..isn't it?

2) CUFF.5.1 - - CUFF.5 if the software identified transcripts that are completely new compared to reference..isn't it?

and finally
3) ENST00000428771 - - CUFF.9
ENST00000304952 - - CUFF.9
ENST00000481869 - - CUFF.9
ENST00000484667 - - CUFF.9
This is the question: if cufflinks doesn't identify new transcripts for this locus (because the 4 related transcripts are known) why the locus (CUFF.9) is not known and the same of the reference? These transcripts have the same coordinate of the reference.

So.....How can I discriminate between novel transcripts (CUFF.5.1 - - CUFF.5) and the new possible isoforms of a known gene?

Thanks