I am using scrublet to remove doublets from my data. I am following the tutorial here:https://github.com/swolock/scrublet/blob/master/examples/scrublet_basics.ipynb
I don't think that the automatic threshold detection worked well for my data. The threshold doublet score is set at 0.26 and the histograms look like this:
I tried different threshold values so I set it to 0.17 and 0.15 to see if the histograms would be improved. I thinks that the 0.15 looks the best but I was wondering if that seems too low.
This is 0.17
And this is 0.15