Using CSI instead of TBI indexing to handle large chromosomes in GTF file

[samuelruizperez]

Hi!

Thank you for developing such a great tool!

I tried to run UROPA 4.0.2 and I got these error messages:

...
...
[DEBUG]   Tabix compress
[E::hts_idx_check_range] Region 536192814..537448215 cannot be stored in a tbi index. Try using a csi index
[WARNING] Indexing failed - the GTF is probably unsorted
[WARNING] Attempting to sort with call: grep -v "^#" /path/uropa/test_feature_subset.gtf | sort -k1,1 -k4,4n > /path/uropa/test_sorted.gtf
[E::hts_idx_check_range] Region 536192814..537448215 cannot be stored in a tbi index. Try using a csi index
[ERROR]   Could not index .gtf-file - please check whether the file has the correct 9-column format.
[ERROR]   Logger lost connection to queue - probably due to an error raised from a child process.

I think the issue is that my GTF has some pretty large chromosomes, and they cannot be stored in a TBI index. Is there an easy way to index as CSI instead of TBI or any other way to handle large chromosomes in the GTF file for UROPA?

Thank you!

[msbentsen]

Hi @samuelruizperez , thank you so much for this issue and the pull request! It would be fine for me to force the use of CSI, but for backwards compatibility, I made a fix for uropa 4.0.3 which automatically checks for the largest coordinates and uses CSI if the chromosomes are too large. It's available on PyPI now - I hope this solves your issue! I will close your pull request but feel free to contribute again in the future, thanks! 🥳

[samuelruizperez]

Hi, @msbentsen:

Thank you for the quick reply and fix! Sounds good to me. 🥳