imsb-uke / nichepca

MIT License
1 stars 1 forks source link

Using pre calculated PCA / normalization #1

Closed pakiessling closed 2 months ago

pakiessling commented 2 months ago

Description of feature

Hi,

I have a dataset of about 2 million cells which I integrated with rapids single cell on gpu. I got much better results by using z-scores for every samples separately followed by harmony. Is there a way to reuse these (lenghty) processing steps for nichepca or do I have to start from counts?

Ty!

dschaub95 commented 2 months ago

Hi,

sorry for the late reply!

I fear you have to start from counts since you need to first run the aggregation before applying PCA and harmony.

Best Darius

pakiessling commented 2 months ago

ok, ty!