CompaRNA - on-line benchmarks of RNA structure prediction methods

FAQ

Metric	CentroidAlifold(seed)		Murlet(20)
MCC	0.761	>	0.666
Average MCC ± 95% Confidence Intervals	0.755 ± 0.084	>	0.679 ± 0.114
Sensitivity	0.623	>	0.548
Positive Predictive Value	0.934	>	0.814
Total TP	283	>	249
Total TN	51960	>	51957
Total FP	38	<	79
Total FP CONTRA	17	=	17
Total FP INCONS	3	<	40
Total FP COMP	18	<	22
Total FN	171	<	205
P-value	2.76499031155e-08

Comparison of performance of CentroidAlifold(seed) and Murlet(20). Positive Predictive Value (PPV) is plotted against sensitivity. Each dot represents a single test of each method. See tables below for raw data (individual counts for CentroidAlifold(seed) and Murlet(20)).
Average Matthews Correlation Coefficients (MCC) with 95% confidence intervals (CIs) were plotted for different RNA families, for which at least 3 members were present in the benchmarking dataset. 'n' denotes the number of MCCs used to calculate the average and CI. See tables below for raw data (individual counts for CentroidAlifold(seed) and Murlet(20)).
Comparison of average Matthews Correlation Coefficients (MCCs) for CentroidAlifold(seed) and Murlet(20). The whiskers correspond to 95% confidence intervals (CIs). 'n' denotes the number of MCCs used to calculate average MCCs and CIs. See tables below for raw data (individual counts for CentroidAlifold(seed) and Murlet(20)).

Total Scores
MCC	0.761
Average MCC ± 95% Confidence Intervals	0.755 ± 0.084
Sensitivity	0.623
Positive Predictive Value	0.934
Nr of predictions	15

RNA Chain	Rfam family	MCC	SENS	PPV	TP	TN	FP	FP CONTRA	FP INCONS	FP COMP	FN
3A2K_C	tRNA links to Rfam database...	0.52	0.27	1.00	6	1102	0	0	0	0	16
3GX2_A	SAM links to Rfam database...	0.92	0.86	1.00	24	1425	1	0	0	1	4
3IVN_B	Purine links to Rfam database...	0.86	0.83	0.90	19	882	2	2	0	0	4
3IZ4_A	tmRNA links to Rfam database...	0.58	0.43	0.77	41	25483	17	12	0	5	54
3JYV_7	tRNA links to Rfam database...	0.50	0.25	1.00	5	1106	1	0	0	1	15
3LA5_A	Purine links to Rfam database...	0.87	0.80	0.95	20	933	1	1	0	0	5
3NPB_A	SAM links to Rfam database...	0.79	0.65	0.96	24	2253	5	0	1	4	13
3O58_3	5_8S_rRNA links to Rfam database...	0.68	0.50	0.92	11	4752	2	1	0	1	11
3PDR_A	ykoK links to Rfam database...	0.93	0.88	0.98	44	4795	3	0	1	2	6
3RKF_A	Purine links to Rfam database...	0.91	0.83	1.00	20	846	0	0	0	0	4
3SD1_A	THF links to Rfam database...	0.85	0.76	0.96	22	1510	1	0	1	0	7
4A1C_2	5_8S_rRNA links to Rfam database...	0.60	0.40	0.89	8	4507	3	1	0	2	12
4AOB_A	SAM links to Rfam database...	0.89	0.79	1.00	23	1414	2	0	0	2	6
4ENB_A	crcB links to Rfam database...	0.72	0.53	1.00	8	464	0	0	0	0	7
4ENC_A	crcB links to Rfam database...	0.73	0.53	1.00	8	488	0	0	0	0	7

Total Scores
MCC	0.666
Average MCC ± 95% Confidence Intervals	0.679 ± 0.114
Sensitivity	0.548
Positive Predictive Value	0.814
Nr of predictions	15

RNA Chain	Rfam family	MCC	SENS	PPV	TP	TN	FP	FP CONTRA	FP INCONS	FP COMP	FN
3A2K_C	tRNA links to Rfam database...	0.88	0.86	0.90	19	1087	2	0	2	0	3
3GX2_A	SAM links to Rfam database...	0.73	0.54	1.00	15	1434	1	0	0	1	13
3IVN_B	Purine links to Rfam database...	0.80	0.65	1.00	15	888	0	0	0	0	8
3IZ4_A	tmRNA links to Rfam database...	0.46	0.33	0.66	31	25489	22	3	13	6	64
3JYV_7	tRNA links to Rfam database...	0.84	0.80	0.89	16	1093	4	0	2	2	4
3LA5_A	Purine links to Rfam database...	0.80	0.64	1.00	16	938	0	0	0	0	9
3NPB_A	SAM links to Rfam database...	0.65	0.43	1.00	16	2262	2	0	0	2	21
3O58_3	5_8S_rRNA links to Rfam database...	0.37	0.32	0.44	7	4748	13	4	5	4	15
3PDR_A	ykoK links to Rfam database...	0.84	0.76	0.93	38	4799	3	1	2	0	12
3RKF_A	Purine links to Rfam database...	0.81	0.67	1.00	16	850	0	0	0	0	8
3SD1_A	THF links to Rfam database...	0.82	0.83	0.83	24	1504	5	4	1	0	5
4A1C_2	5_8S_rRNA links to Rfam database...	0.22	0.25	0.20	5	4491	24	5	15	4	15
4AOB_A	SAM links to Rfam database...	0.87	0.76	1.00	22	1415	3	0	0	3	7
4ENB_A	crcB links to Rfam database...	0.51	0.27	1.00	4	468	0	0	0	0	11
4ENC_A	crcB links to Rfam database...	0.57	0.33	1.00	5	491	0	0	0	0	10

Matthews Correlation Coeffient, Sensitivity and Positive Predictive Value have been calculated based on the paper by Gardener & Giegerich, 2004.