Low regulon recovery in pySCENIC ctx for two butterfly scRNA-seq datasets

[francicco]

Hi,

I’m quite new to this type of analysis, and I’m encountering some issues while running the SCENIC pipeline on two single-cell RNA-seq datasets from closely related butterfly species.

I generated two cisTarget databases using Drosophila orthologs, considering 5 kb upstream and 2 kb downstream of the TSS locations for each species. The earlier steps of the pipeline ran successfully, but I’m having trouble with the pyscenic ctx step.
• For the first species, I obtained 60 regulons out of 307 TFs, which already feels a bit low.
• For the second species (which should be slightly better annotated and processed using the same approach), I got 0 regulons out of 318 TFs.

I’d really appreciate any advice on what might be going wrong or suggestions for improving the regulon recovery — for example, whether this might relate to database composition, gene name matching, or parameter choices.

Thanks a lot for your help!
Francesco