Issue with ERCC spike-ins in the GTF when using run_smartseq2 about velocyto.py HOT 3 CLOSED

Sorry for the late reply.
velocyto is looking for the following standard entry in the gtf files:

transcript_id, transcript_name, gene_id, gene_name, exon_number

This info should be standerd in a gtf file formated according GENCODE specification. Here you are missing transcript_name gene_name and exon_number.

However I see the issue here those entities don't make sense for the ERCC spikes... I could easily catch this error and return some default not informative value like

transcript_id="NoId", transcript_name="NoName", gene_id="NoId", gene_name="NoName", exon_number="1"

However I am afraid that this sometimes will lead users that are using incorrectly formatted gtf files to some weird outputs.

I think the best solution is to still throw an error and return a more informative error message. (Even though here the second last line was kind of clear that the regex_trname was failing, but I can try to be even more explicit).

For you the best solution is to just add those entries in the gtf

from velocyto.py.

Note that now velocyto will be a little more forgiving and if transcript_name or exon_number are not specified no error will be thrown.

from velocyto.py.

Velocyto now also prints the line of the gtf that cause the error, helping the user to debug. I still think that too much "secret patching" of the gtf data where the user is not aware of what is going on, should be avoided, because it might generate anomalies difficult to trace back.

from velocyto.py.