Has stochasticity been removed?

[lucygarner]

Hi @s-aibar, @cbravo93,

I have a custom pipeline that runs pySCENIC 100 times and then aggregates the results to define high confidence regulons as discussed here (https://pyscenic.readthedocs.io/en/latest/faq.html).

However, I have just downloaded the current version of pySCENIC (0.12.1) and run my pipeline, and all of the runs give identical results. So the stochasticity seems to have been removed. I use arboreto_with_multiprocessing.py (https://github.com/aertslab/pySCENIC/blob/master/src/pyscenic/cli/arboreto_with_multiprocessing.py) for gene regulatory network inference.

Has the stochasticity been removed for a reason?

Best wishes,
Lucy

[lucygarner]

Apologies, I realised that I added a seed to the first step, thinking that this was a good idea. But that defeats the point of allowing the stochasticity!