Fig. 4

Boxplot showing the values of the calculated metrics using CCA, ComBat, PCA, MMDAE followed by Harmony with dimension (256, 32), PCA followed by BBKNN, and PCA followed by Harmony as well as by the best AE for each of the tested dimension (H, L), analysing the PBMC datasets. a AMII achieved by the different strategies. b AMII achieved by the best AE for each of the tested dimension. c FMS achieved by the different strategies. d FMS achieved by the best AE for each of the tested dimension. e HS achieved by the different strategies. f HS achieved by the best AE for each of the tested dimension. g CS achieved by the different strategies. h CS achieved by the best AE for each of the tested dimension. i VM achieved by the different strategies. j VM achieved by the best AE for each of the tested dimension. p-value \(\le 0.0001\) (****); \(0.0001<p\)-value \(\le 0.001\) (***); \(0.001<p\)-value \(\le 0.01\) (**); \(0.01<p\)-value \(\le 0.05\) (*); p-value \(>0.05\) (ns)