search

mortazavilab / cerberus

14 stars 1 forks source link

Cannot compute TSSes or TESes without strand info #6

Open EmilVorsteveld opened 2 months ago

EmilVorsteveld commented 2 months ago

Using GTF files from GENCODE, I get the following error:

cerberus gtf_to_bed --gtf gencode.v45.annotation.gtf -o gencode.v45.annotation.out --mode tss

Exception: Cannot compute TSSes or TESes without strand info. Perhaps use extend() or subsequence() or spliced_subsequence() instead?

fairliereese commented 2 months ago

Without digging into the issue too much, I have a feeling this might be caused by strange PyRanges stranded vs. unstranded behavior that I've been encountering lately (see this issue). Either way, I recommend installing my version of PyRanges for use with Cerberus. Try this and let me know if it solves your issue!

EmilVorsteveld commented 2 months ago

This indeed solves this issue, thanks. Unfortunately, for me, the agg_ics function throws the following error:
File "/Users/emilvorsteveld/miniforge3/envs/cerberus/lib/python3.9/site-packages/pandas/core/internals/base.py", line 57, in _validate_set_axis raise ValueError( ValueError: Length mismatch: Expected axis has 4 elements, new values have 3 elements

I have tried older pandas versions (1.3.0 and 1.5.3), but not luck...

fairliereese commented 2 months ago

Hi, sorry for the delay. Can you please attach the complete error message stack trace so I can see what line in Cerberus threw the error?

Thanks!