search

weberlab-hhu / Helixer

Using Deep Learning to predict gene annotations
GNU General Public License v3.0
140 stars 20 forks source link

Helixer.py produces empty GFF file for very small Fasta Input #83

Closed soi closed 1 year ago

soi commented 2 years ago

test.fa.txt

Helixer.py produced an 0 Byte GFF file as output for this input despite calling the process successful on the cli. Larger input files worked.

alisandra commented 1 year ago

Thanks, I finally took a look, and I got output with the provided test.fa.txt, (using either land_plant_v0.3_a_0100.h5 or vertebrate_v0.3_m_0100.h5).

Did you maybe fix it but not close the issue? Else, I'd hazard a guess that maybe the model used before simply hadn't predicted a gene there.

In any case, HelixerPost should now include the pragmas for sequence start, etc, regardless of a sequence without a gene; so this sort of thing would now be a less cryptic error.

Closing now in any case, do reopen if I'm missing something.