Skip to main content
Figure 4 | Genome Biology

Figure 4

From: Inferring mouse gene functions from genomic-scale data using a combined functional network/classification strategy

Figure 4

Biological process GO terms showing highly divergent predictability by the network and the classifier ensemble. Specific functional biases of biological process (BP) terms between the network and classifier ensemble are highlighted by plotting the difference between AUCnetwork_slim and AUCclassifier for GO terms showing the largest ΔAUC = max(AUCnetwork, 0.5) - max(AUCclassifier, 0.5). The GO terms with ancestor-descendant relationships were merged to the ancestor term to remove redundancy. The Gene Ontology (GO) terms with |ΔAUC| > 0.15 were plotted. As the overall performance was better in the network than in the classifier, most GO terms show positive ΔAUC. For brevity, the maximum number of terms was set to 15 and 65 for negative and positive ΔAUCs, respectively. AUC, area under the receiver operating characteristic.

Back to article page