Cuffdiff reporting same locus several times &/or more than one gene per transcript

[Marta123]

Hi all,

I am doing RNA-Seq of pig (Sus scrofa 10.2) and I got some weird results in cuffdiff output and I wonder if someone did ever find them. (I am using the lastest versions of TopHat and Cuffflinks as well as the latest genome and gtf files from ensembl).

1. More than one gene_id reported in one transcript

2. More than one transcript with exactly the same locus

e.g Some of the output that I obtained in genes.fpkm_tracking look like this:

XLOC_000006 ENSSSCG00000004008,ENSSSCG00000025112,ENSSSCG00000030155 TSS10,TSS11,TSS12,TSS13,TSS6,TSS7,TSS8,TSS9 1:935011-1369891
XLOC_000007 ENSSSCG00000021794 TSS14 1:935011-1369891

XLOC_000026 ENSSSCG00000026205,PARK2 TSS45,TSS46,TSS47,TSS48,TSS49 1:7195765-8262315
XLOC_000027 - TSS50 1:7195765-8262315
XLOC_000028 - TSS51 1:7195765-8262315

I obtained up to 246 transcripts with > 1 gene annotated per transcript.
Is it a program bug? Running edgeR parallely did not report this output... So I don't think they are fusion genes...

Any idea?