Skip to main content


Figure 2 | BMC Bioinformatics

Figure 2

From: A domain-centric solution to functional genomics via dcGO Predictor

Figure 2

The performance of dcGO Predictor in the CAFA experiment. To evaluate function prediction, true prediction rate (precision: TP/[TP+FP]) and true positive rate (recall: TP/[TP+FN]), both as a function of the predictive score (see Methods) are plotted as a precision-recall (PR-RC) curve. (A) PR-RC curves for each of 7 sets of eukaryotic sequence targets, separately calculated using (top panel) the Molecular Function (MF) and (bottom panel) Biological Process (BP) ontologies. These predictions are based on GO annotations of both domains and supra-domains. (B) Comparisons of prediction using both domains and supra-domains against that using domains. PR-RC curves plotted here are for all eukaryotic sets as a whole. (C) Comparisons of PR-RC curves between eukaryotic sets and prokaryotic sets.

Back to article page