Open wangjiawen2013 opened 2 years ago
Have you tried to use highly variable features only?
No, I think it's better to use the whole transcriptome with minimized filtering according to the cell system paper. Can we utilize parallel processing to reduce the time ?
Will hotspot support parallel processing ?
Will hotspot support parallel processing ?
For my data hs.compute_local_correlations() and hs.compute_autocorrelations() are a little bit slow but I can set jobs=20 for parallel processing. It took about 40 mins on 8000 cells * 5000 genes. create_modules() only took one or two seconds. I do not think subsetting genes/cells makes a big difference.
Hi, hotspot create_modules runs slowly, it tooks me about 4 days on 50,000 cells how to speed up ?