Figure 3 from manuscript Sparsely-Connected Autoencoder (SCA) for single cell RNAseq data mining

Dataset used to generate figure 3: COMET analysis of latent space frequency matrix. A) Set of 4 genes (PAX5, NFAT5, RFXANK and CHD4) characterizing cluster 1 (B-cells) (figure3/setA/Results/setATF_SIMLR/5/outputvis/cluster_1_discrete_quads/rank_1.png). B) Set of 4 genes (CEBPA, KHSRP negation, CEBPB, CREBBP) characterizing cluster 2 (Monocytes) (figure3/setA/Results/setATF_SIMLR/5/outputvis/cluster_2_discrete_quads/rank_1.png). Dots are cells, blue and red colour indicates respectively false and true positives (for more information on this type of visualization of cluster specific markers see the COMET paper). C) Rank 1 NK signature specifically characterizing cluster 4 (NK) (figure3/setA/Results/setAIS2_KMEAN/5/outputvis/cluster_4_singleton/rank_1.png). D) Rank 2 Asthma KEGG negation, specifically characterizing Naïve T-cells (Cluster 5) (figure3/setA/Results/setAIS2_KMEAN/5/outputvis/cluster_5_singleton/rank_2.png).