When the number of cell clusters is large, the memory may be insufficient for calculating flux. Are there any solutions to this problem?

[MOSAIFENG]

The following is the code I used.

library(Seurat)
saveRDS(seurat_obj, file="XXXX.rds")

library(METAFlux)
data("human_blood")

seurat_obj <- readRDS("/home/saifeng/cross-talk/METAFlux/scanpy4_seurat_obj.rds")
mean_exp=calculate_avg_exp(myseurat = seurat_obj,myident = 'cell_type',n_bootstrap=3,seed=1)

scores <- calculate_reaction_score(data = mean_exp)

cell_fractions <- round(table(seurat_obj$cell_type) / nrow([email protected]), 3)
flux <- compute_sc_flux(num_cell = 24, fraction = cell_fractions, fluxscore = scores, medium = human_blood)
Preparing for TME S matrix.....
Error: cannot allocate vector of size 742.2 Gb

[seigfried]

flux=compute_sc_flux(num_cell = nrow(test),
+                      fraction = test$Fraction,
+                      fluxscore=scores,
+                      medium = cell_medium)

Preparing for TME S matrix.....
Error: cannot allocate vector of size 1316.0 Gb

In my case I had an object of 25337 cells only, I had 16 celltypes and 2 conditions hence 32 columns in scores with a testing bootstrap of 1. How can I adjust memory and run this? Will try running 1 celltype at a time and see if it works