Since gff files have specific parent/child structure, they are a natural fit for representing splicing events as isoforms. Additionally, the gff files created by outrigger can be combined with the original gtf/gff files for ease of biological annotation, e.g. protein translation.
The IDs in the gff files would have to be exon-based to be unique: exon1@exon2@exon3 so in events.csv, we would have to add a exon_based_event_id column, in addition to the current event_id which is junction based (e.g. for SE isoform1=junction13|isoform2=junction12@exon2@junction23)
Since gff files have specific parent/child structure, they are a natural fit for representing splicing events as isoforms. Additionally, the gff files created by outrigger can be combined with the original gtf/gff files for ease of biological annotation, e.g. protein translation.
The IDs in the gff files would have to be exon-based to be unique:
exon1@exon2@exon3
so inevents.csv
, we would have to add aexon_based_event_id
column, in addition to the currentevent_id
which is junction based (e.g. for SEisoform1=junction13|isoform2=junction12@exon2@junction23
)