Skip to main content
. 2021 Jun 8;22:309. doi: 10.1186/s12859-021-04150-3

Fig. 5.

Fig. 5

UMAP visualisation showing the sample alignment performed by Harmony into the latent space obtained by MMDAE with dimension (256, 32) for the PBMC datasets a, by GMMMD with dimension (256, 32) for the PIC datasets b, and by MMDVAE with dimension (256, 64) for the MCA datasets c