This is the code I am using for perform cluster based size factor normalization on a single cell RNA seq data with around 7 - 9 lac cells

set.seed(123) param <- SnowParam(workers = 5, type = "SOCK", progressbar = TRUE, RNGseed = 123) clusters <- quickCluster(sce_object, block=cut(seq_len(ncol(counts(sce_object))), 25), block.BPPARAM = param) sce_object <- computeSumFactors(sce_object, clusters = clusters, BPPARAM = param)

However, its taking too much time and ultimately getting "Killed" due to memory constraint. Could you please suggest some alternative or do you think random downsampling of cells is the only solution?

Many thanks and regards Amlan.

The obvious solution is to use cloud computing, which is likely cheaper than getting more RAM for your computer.