Table 2.

Summary of Classification Performance for sPLS-DA and PLS-DA Classifiers Trained on Full 26-Sample Training Set as Determined on a 14-Sample Testing Set (7AR and 7NR)

	Classifier	Number of selected variables	Classification Error rate (CE)	Sensitivity	Specificity	Area under curve
Genomics	sPLS-DA	90	0.14	0.71	1	0.90
	PLS-DA	27,306 (all)	0.21	0.57	1	0.82
Proteomics	sPLS-DA	21	0.29	0.57	0.86	0.76
	PLS-DA	133 (all)	0.43	0.43	0.71	0.55