Skip to main content
Fig. 8 | Journal of Translational Medicine

Fig. 8

From: Improving knowledge on the activation of bone marrow fibroblasts in MGUS and MM disease through the automatic extraction of genes via a nonnegative matrix factorization approach on gene expression profiles

Fig. 8

a Heatmap plot of basis matrix extracted from the gene expression profile data matrix. This matrix has been obtained concatenating expression data from 10 patients with MM condition and 8 patients with MGUS condition. Five metagenes were automatically extracted with a rank of the factorization = 5. Values in each column of the basis matrix have been normalized by row and sorted to show higher values on the bottom and lower values on the top of the heatmap. The first metagene was automatically identified as the most representative metagene of the whole dataset of the two conditions, since it presented the highest number of relevant values. b Density level of the coefficient matrix for each patient in the first metagene. Values in the coefficient matrix have been normalized by column to a clearer result representation. The MM and MGUS conditions are marked by circle and star markers, respectively. Higher values represent a greater influence of genes in the metagene on the corresponding patient

Back to article page