Markup | Genome Research

Table 4.

Classification Quality During Cross-Validation

Process	AUC	SE	Sensitivity	Specificity
Ion homeostasis	1.00	0.00	1.00	1.00
Protein targeting	0.99	0.03	1.00	0.98
Blood coagulation	0.96	0.08	0.96	0.99
DNA metabolism	0.94	0.09	0.94	0.93
Intracellular signaling cascade	0.94	0.06	0.92	0.94
Energy pathways	0.93	0.12	0.89	0.99
Cell cycle	0.93	0.04	0.93	0.86
Oncogenesis	0.92	0.11	0.94	0.93
Circulation	0.91	0.11	0.87	0.95
Cell death	0.90	0.10	0.85	0.90
Developmental processes	0.90	0.07	0.91	0.90
Transcription	0.88	0.11	0.84	0.82
Defense (immune) response	0.88	0.05	0.88	0.91
Cell adhesion	0.87	0.09	0.85	0.91
Stress response	0.86	0.15	0.87	0.89
Protein metabolism and modification	0.85	0.10	0.83	0.86
Cell motility	0.84	0.11	0.83	0.89
Cell surface receptor-linked signal transduction	0.82	0.15	0.79	0.84
Lipid metabolism	0.81	0.14	0.77	0.85
Transport	0.79	0.17	0.72	0.84
Cell organization and biogenesis	0.79	0.11	0.76	0.91
Cell proliferation	0.79	0.06	0.76	0.77
Amino acid and derivative metabolism	0.69	0.06	0.29	0.98
Average	0.88	0.09	0.84	0.91

[i] Tenfold cross-validation estimates of the area under the ROC curve (AUC), standard error (SE) for AUC, and sensitivity and specificity for each of the 23 biological processes. Sensitivity is TP/(TP + FN) where TP (true positives) is the number of genes classified and annotated to the process and FN (false negatives) is the number of genes annotated but not classified to it. Specificity is TN/(TN + FP), where TN (true negatives) is the number of genes neither annotated nor classified to the process and FP (false positives) is the number of genes classified but not annotated to it.