Dinucleotide controlled null models for comparative RNA gene prediction

Table 1 z-scores and classification performance

		RNAz			AlifoldZ			SISSIz (mono)			SISSIz (di)
Data type	N	z	S_0.01	S_0.05	z	S_0.01	S_0.05	z	S_0.01	S_0.05	z	S_0.01	S_0.05
5S rRNA	368	n/a	0.77	0.98	-6.72	0.84	0.98	-6.35	0.86	0.98	-6.35	0.93	1.00
tRNA	382	n/a	0.74	0.98	-6.29	0.75	0.98	-6.24	0.74	0.98	-5.86	0.88	0.99
U2 snRNA	458	n/a	0.76	1.00	-7.17	0.89	0.99	-5.92	0.84	0.97	-5.22	0.93	0.99
U3 snRNA	377	n/a	0.52	0.92	-5.11	0.74	0.86	-4.47	0.69	0.83	-4.23	0.76	0.86
U5 snRNA	424	n/a	0.90	0.96	-5.61	0.77	0.96	-5.10	0.69	0.89	-4.43	0.76	0.91
Hammerhead	499	n/a	0.78	1.00	-6.68	0.85	1.00	-6.67	0.90	1.00	-6.66	0.99	1.00
Group II intron	480	n/a	0.68	0.82	-6.58	0.74	0.81	-6.77	0.72	0.81	-6.29	0.77	0.82
micro RNA precursor	571	n/a	0.75	1.00	-8.89	1.00	1.00	-8.84	1.00	1.00	-7.58	1.00	1.00
Total of all classes	3559	n/a	0.80	0.96	-6.75	0.87	0.95	-6.43	0.85	0.94	-5.93	0.90	0.95
Genomic background	3559	n/a	n/a	n/a	-0.44	n/a	n/a	-0.58	n/a	n/a	-0.15	n/a	n/a

S_0.01, S_0.05 ... Sensitivity at a false positive rate of 0.01 and 0.05, respectively. The highest values among the four different methods are shown in bold letters.