A clustering-based approach for efficient identification of microRNA combinatorial biomarkers

BMC Genomics

Table 8 Accuracies of different feature subsets^a

Feature subset	Accuracy measure	GSE22220				GSE40525
		T-test	FDA	SNR	Center	T-test	FDA	SNR	Center
All^a	Sensitivity	0.874	0.882	0.921	0.906	0.962	0.962	0.962	0.942
	Specificity	0.804	0.794	0.770	0.762	0.806	0.781	0.806	0.817
	TA	0.792	0.787	0.783	0.768	0.865	0.846	0.865	0.865
Pair	Sensitivity	0.969	0.969	0.929	0.984	0.923	0.923	0.885	0.846
	Specificity	0.750	0.750	0.756	0.714	0.923	0.923	0.920	0.917
	TA	0.783	0.783	0.773	0.749	0.923	0.923	0.904	0.885
Triple	Sensitivity	0.976	0.976	0.850	0.984	0.942	0.942	0.942	0.904
	Specificity	0.775	0.775	0.812	0.714	0.925	0.925	0.925	0.922
	TA	0.812	0.812	0.787	0.749	0.933	0.933	0.933	0.913
Quadruple	Sensitivity	0.906	0.906	0.906	0.890	0.942	0.942	0.887	0.923
	Specificity	0.833	0.833	0.821	0.819	0.961	0.961	0.940	0.960
	TA	0.831	0.831	0.821	0.812	0.952	0.952	0.915	0.942

^aAll: the full set of representative miRNAs selected from clusters Sensitivity = TP/(TP+FN), Specificity = TN/(FP+TN) TA: total accuracy

ISSN: 1471-2164