Skip to main content
Fig. 2 | BMC Bioinformatics

Fig. 2

From: Deconvolution of autoencoders to learn biological regulatory modules from single cell mRNA sequencing data

Fig. 2

Impact of hematopoietic signatures on the representation layer of our autoencoder

Impact of hematopoietic signatures (rows) on the representation layer (columns) of the autoencoder trained on PBMC data. The hematopoietic signatures are derived from the DMAP study. To enhance visual clarity, only the high impact pathways are visualized. The top-left heat map corresponds to all the cells. The rest of the heat maps correspond to a summary of cells in each cellular population of the study, as clustered and labeled by Seurat. Row names correspond to cell types categories, or to DMAP labels for sub-classification, where TCELL8A is CD4 + CD45RA-CD62L+ T-cells and BCELL4 is CD19 + lgD-CD27+ B-cells, respectively. Not shown are Seurat predicted clusters on Megakaryocytes cells (< 1% in human PBMC)

Back to article page