Skip to main content
. 2023 Nov 22;24:440. doi: 10.1186/s12859-023-05569-6

Table 3.

Time (in s) and memory (in MiB) for execution of Scanpy, SC3 and Seurat methods in several single cell banks

Method
Data Number of Number of Scanpy SC3 Seurat
set cells genes Time Memory Time Memory Time Memory
Wang 71,032 100 81.97 9194.92 (1) 123.75 483.89
500 100.69 9541.20 (1) 151.63 1238.48
4000 229.94 11786.49 2907.28 8334.96 299.47 5828.26
all 178.63 11572.80 6202.98 40560.79 717.55 27332.94
Garcia 100,307 100 117.89 13616.54 (1) 202.43 476.88
500 142.09 14174.46 (1) 204.96 940.13
4000 338.65 17515.35 3047.56 8411.55 347.43 5397.71
all 340.17 12055.15 6050.07 51085.32 1198.42 32640.66
Fonseca 118,144 100 120.84 8014.76 (1) 191.96 448.31
500 159.91 8944.68 (1) 252.76 1002.43
4000 388.47 12011.64 2606.85 9017.33 415.30 4996.09
all 178.63 11572.80 4752.91 73529.04 1678.49 31231.88
Merge 289,483 100 638.00 30738.78 (1) 631.02 950.68
500 490.14 32033.64 (1) 706.44 2266.57
4000 1111.51 40586.07 2960.67 21616.60 1017.365 13205.54
all 1454.21 41961.79 (2) (2)

SC3 was applied with 64 threads; Scanpy and Seurat do not allow the number of threads to be chosen

(1): Application of SC3 with 100 and 500 genes caused a program error

(2): Program crashes without error messages due to memory exhaustion on a 256 GiB machine