Figure 4From: Simultaneous identification of robust synergistic subnetwork markers for effective cancer prognosisReproducibility of various subnetwork and pathway markers. In order to evaluate the reproducibility of various modular markers, we used the first dataset to identify potential markers and select the optimal set of features and the second dataset to train the classifier (using the selected features) and evaluate its performance. Average classification performance is shown when the markers were selected based on (a) the USA dataset, (b) the Netherlands dataset, (c) the Belgium dataset, and (d) the Sweden dataset.Back to article page