Vipul Singhal

Hi Alwash, have you tried using hightly variable genes (2000 HVGs in Seurat)? Another optimization to try is to first do HVG to 2000 genes, then further reduce to 100...

Also: 1. construct the BANKSY matrix separately for each sample, merge them, and run PCA on the merged matrix. 2. See this: https://github.com/prabhakarlab/Banksy_py/issues/12#issuecomment-2268114768 Let me know how it goes! Best,...

Thanks for the suggestion! will take a look at this soon.

You need to run things in multi sample mode. See [here](https://github.com/jleechung/seurat-wrappers/blob/feat-aft/docs/banksy.md#multi-sample-analysis) for a canonical workflow if using Seurat. If not using Seurat, use [this](https://prabhakarlab.github.io/Banksy/articles/multi-sample.html) vignette. If interested, #61 is related....

No there should not be any issues with 0.8, but if you are doing spot based technologies like Visium, where each spot is already an average of many cells, 0.2...