Fig. 2
NanoString gene expression profiling of DDLPS versus non-malignant samples. (A) Heatmap representing the mRNA expression profiles of all samples. Differential expression analysis was performed using the Limma pipeline, which applies a moderated t-statistic to identify differentially expressed genes. Light green and pale blue annotations on the right represent, respectively, genes differentially expressed in DDLPS versus non-malignant tissues. The majority of differentially expressed genes in DDLPS versus non-malignant tissues are enriched in two different clusters with high within-group internal similarity. Columns represent samples and rows represent genes. Columns are split according to their clinical behavior, and, within each group, columns are sorted according to their similarity. Genes are sorted according to their similarity. Hierarchical clustering was calculated using Euclidean distance. The expression is by-row normalized to better show between-samples differences. (B) Volcano plot of differential gene expression comparing DDLPS versus non-malignant samples. Red and blue dots: significantly (abs(log2(FC)) > 1 and adjusted p value < 0.1) upregulated or downregulated genes; black dots: genes with non-significant (abs(log2(FC)) < 1 or adjusted p value > 0.1) alteration; x axis shows base-2 logarithm fold change; y axis shows negative base-10 logarithm of unadjusted p values; FC: fold change.
