Fig. 2

Classification efficiency for the selected proteins. a Heatmap with hierarchical clustering. In the heatmap, the columns and rows represent the 60 EV samples belonging to the nine tumor types marked with different colors and the 172 proteins, respectively. Both the columns and rows are clustered. Dendrogram branches ending in a square indicate the elements to be included in the Train set. b t-SNE plot of the selected 172 proteins. The dots with different colors represent the 60 individual EV samples belonging to the nine tumor types. In the plot, the color gradient indicates the dot density. c Confusion matrix of the classification results using the selected proteins on the Test set. Each row of the matrices represents the instances in an actual class while each column represents the instances in a predicted class. Diagonally, the percentage of the correct classification is shown in blue. The percentage of errors is indicated in red