Understanding the selection of most variable genome bin

[wangmeijiao]

Hi Kai and all,
Here I raised some questions to find help from you. I understand snapatac2 will choose some most variable features (genomic bins) to perform spectral clustering (stored in var: 'selected', by pp.select_feature). Then, 1) what is the selection strategy? 2) if I combine multiple processed snap object as anndataset, will this var['selected'] be sum as a final boolean variable. 3) Or should I rerun this step (pp.select_feature) for the merged object? Sorry for the naive question and I hope you can find and answer my question in this discussion thread.

And, to avoid open another thread, here is a quick question: will the function module ex.export_bigwig() perform normalization for the aggregated Tn5 insertion counts? Because this is important for difference comparison of DARs!

Hoping these questions will help others too!

Yours,
Meijiao