powellgenomicslab / Nebulosa

R package to visualize gene expression data based on weighted kernel density estimation
89 stars 8 forks source link

What is the optimal threshold value of joint density? #7

Open yuGithuuub opened 3 years ago

yuGithuuub commented 3 years ago

Dear Nebulosa group: Thanks for your wonderful work in single cell visualization! I've found that for well expressed markers, there's great performance. But for some obviously wrong combinations, it will show a bad result,because of the joint density value. What is the best threshold? Can I adjust the maximum value manually?

Rplot1 Rplot2

joseah commented 3 years ago

Hi @yuGithuuub, thanks for your observation. Most densities are non-zero, resulting in low joint densities values when markers are not "co-expressed". We are currently working on improving the visualization for this scenario. Any recommendations or feedback are very well welcome.

yuGithuuub commented 3 years ago

Looking forward to the improved version!

BioMedInvestorNL commented 1 year ago

Hi. Do you have an update on this? On how to set thresholds for the joint densities?

best