search

randel / HiDecon

Hierarchical cellular deconvolution
2 stars 1 forks source link

Speed optimization #1

Closed rrydbirk closed 1 year ago

rrydbirk commented 1 year ago

Thank you for an interesting tool!

I took the liberty of conducting some speed optimization. Specifically, I added the possibility for multithreading for the scran::findMarkers function which can be really slow. Also, for select_HiDecon() instead of doing repeated log2 transformations for "ref", since this transformation can be heavy I moved this up to do it once in the beginning of the function.

BR

Rasmus

rrydbirk commented 1 year ago

Hi @Penghuihuang2000

I hope you'll consider some of my additions to HiDecon as it sped up my calculations significantly.

BR

Rasmus

Penghuihuang2000 commented 1 year ago

Hi Rasmus,

Of course! Thank you for your effort. I'll merge after double checking all the details are working perfectly.

Best, Penghui