Table 3.

Accuracy on the high-quality structural alignment dataset.

	ROC score (stdev)
Family	Profile BPLA kernel	Profile LA kernel	Profile stem kernel	RNAz
C/D snoRNA	0.95 (0.02)	0.79 (0.04)	0.80 (0.02)	0.78 (0.03)
H/ACA snoRNA	0.97 (0.02)	0.65 (0.20)	0.89 (0.04)	0.95 (0.03)
miRNA precursor	0.97 (0.01)	0.69 (0.02)	0.92 (0.01)	0.96 (0.01)
Riboswitch	0.92 (0.05)	0.41 (0.23)	0.77 (0.05)	0.97 (0.02)
tRNA	1.00 (0.00)	0.88 (0.03)	0.95 (0.02)	0.96 (0.02)
Average	0.96 (0.02)	0.69 (0.10)	0.86 (0.03)	0.92 (0.02)

Family: name of the target ncRNA family. ROC score: area under the ROC curve. Profile BPLA kernel is compared to the other prediction methods which also utilize the profile information of alignment data: Profile LA kernel, Profile stem kernel, and RNAz.