search

randel / HiDecon

Hierarchical cellular deconvolution
2 stars 1 forks source link

`normalized` in HiDecon #4

Closed KaWingLee9 closed 3 months ago

KaWingLee9 commented 5 months ago

Since normalized needs to be specified in select_HiDecon, it needn't been specified in HiDecon. Any restriction of data input when using HiDecon?

Penghuihuang2000 commented 5 months ago

Hi Kawing, Thanks for your interest in HiDecon. First of all, compared with HiDecon function, I will always suggest using _selectHiDecon since it can possibly provide a better penalty parameter. Second, in HiDecon function, both "bulk" and "ref" input should be normalized data (the same normalization method). What was used in original HiDecon paper is CPM (count per million). There's no further restriction. Feel free to ask if there's any question. Penghui

KaWingLee9 commented 5 months ago

Oh, I see. Thank you for your reply. You really do me a favor.