Skip to main content
. 2022 May 7;50(9):4877–4899. doi: 10.1093/nar/gkac317

Table 2.

Shape of imputed matrix of different imputation methods

Methods D1 D2 D3 D4 D5 D6 D7 D8
scImpute 500*961 500*961 500*962 500*961 500*964 500*964 350*16383 758*16383
SAVER 500*961 500*961 500*962 500*961 500*964 500*964 350*16383 758*16383
MAGIC 500*961 500*961 500*962 500*961 500*964 500*964 350*16383 758*16383
DCA 500*961 500*961 500*962 500*961 500*964 500*964
ALRA 500*961 500*961 500*962 500*961 500*964 500*964 350*16383 758*16383
DrImpute 500*933 500*939 500*939 500*939 500*941 500*941 350*16604 758*16842
DeepImpute 500*961 500*961 500*962 500*961 500*964 500*964 350*16383 758*16383
scTSSR 500*888 500*903 500*907 500*906 500*909 500*911 350*13711 758*13037
scNPF 350*16383 758*16383
AutoImpute 500*883 500*891 500*893 500*898 500*899 500*900 350*1000 758*1000
scIGANs 500*961 500*961 500*962 500*961 500*964 500*964 350*16383 758*16383
scGNN 500*938 500*945 500*943 500*945 500*945 500*946 350*2000

* In this table, the shape is shown as gene*cell. For example, 500*961 means a matrix with 500 cells and 961 genes. – means this imputation method did not run on this dataset. D1–6 are simulated datasets in Dataset 1, D7 represents Dataset 2, and D8 is corresponding to Dataset 3.