CompaRNA - on-line benchmarks of RNA structure prediction methods

FAQ

Metric	CentroidHomfold‑LAST		ProbKnot
MCC	0.555	>	0.493
Average MCC ± 95% Confidence Intervals	0.490 ± 0.154	>	0.449 ± 0.138
Sensitivity	0.397	<	0.439
Positive Predictive Value	0.784	>	0.563
Total TP	207	<	229
Total TN	61573	>	61430
Total FP	66	<	198
Total FP CONTRA	5	<	25
Total FP INCONS	52	<	153
Total FP COMP	9	<	20
Total FN	315	>	293
P-value	3.13260312532e-08

Comparison of performance of CentroidHomfold-LAST and ProbKnot. 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 CentroidHomfold‑LAST and ProbKnot).
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 CentroidHomfold‑LAST and ProbKnot).
Comparison of average Matthews Correlation Coefficients (MCCs) for CentroidHomfold-LAST and ProbKnot. 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 CentroidHomfold‑LAST and ProbKnot).

Total Scores
MCC	0.555
Average MCC ± 95% Confidence Intervals	0.490 ± 0.154
Sensitivity	0.397
Positive Predictive Value	0.784
Nr of predictions	18

RNA Chain	Rfam family	MCC	SENS	PPV	TP	TN	FP	FP CONTRA	FP INCONS	FP COMP	FN
2LA5_A	-	0.46	0.26	0.83	5	624	1	0	1	0	14
2LC8_A	GP_knot1 links to Rfam database...	-0.01	0.00	0.00	0	1529	11	0	11	0	20
2LKR_A	-	0.00	0.00	0.00	0	6101	4	0	4	0	39
3AMU_B	tRNA links to Rfam database...	0.82	0.70	0.95	19	2983	3	0	1	2	8
3J0L_2	-	0.49	0.24	1.00	8	6208	0	0	0	0	25
3J0L_7	-	0.60	0.41	0.88	7	1217	1	0	1	0	10
3J0L_g	-	0.00	0.00	0.00	0	465	0	0	0	0	4
3J0L_a	-	-0.01	0.00	0.00	0	1125	3	0	3	0	16
3J16_L	tRNA links to Rfam database...	0.75	0.57	1.00	17	2758	0	0	0	0	13
3RKF_A	Purine links to Rfam database...	0.72	0.53	1.00	18	2193	0	0	0	0	16
3SD1_A	THF links to Rfam database...	0.64	0.45	0.90	19	3895	2	0	2	0	23
3U4M_B	-	0.58	0.43	0.80	16	3140	4	0	4	0	21
3UZL_B	tRNA links to Rfam database...	0.72	0.54	0.95	20	3549	1	0	1	0	17
4A1C_2	5_8S_rRNA links to Rfam database...	0.16	0.15	0.18	5	11753	29	5	18	6	28
4A1C_3	5S_rRNA links to Rfam database...	0.68	0.56	0.83	30	7104	6	0	6	0	24
4AOB_A	SAM links to Rfam database...	0.71	0.50	1.00	21	4350	1	0	0	1	21
4ENB_A	crcB links to Rfam database...	0.76	0.58	1.00	11	1264	0	0	0	0	8
4ENC_A	crcB links to Rfam database...	0.76	0.58	1.00	11	1315	0	0	0	0	8

Total Scores
MCC	0.493
Average MCC ± 95% Confidence Intervals	0.449 ± 0.138
Sensitivity	0.439
Positive Predictive Value	0.563
Nr of predictions	18

RNA Chain	Rfam family	MCC	SENS	PPV	TP	TN	FP	FP CONTRA	FP INCONS	FP COMP	FN
2LA5_A	-	0.45	0.32	0.67	6	621	3	0	3	0	13
2LC8_A	GP_knot1 links to Rfam database...	-0.01	0.00	0.00	0	1525	15	2	13	0	20
2LKR_A	-	0.85	0.77	0.94	30	6073	4	0	2	2	9
3AMU_B	tRNA links to Rfam database...	0.65	0.59	0.73	16	2981	8	0	6	2	11
3J0L_2	-	0.27	0.27	0.28	9	6184	26	2	21	3	24
3J0L_7	-	-0.01	0.00	0.00	0	1212	13	0	13	0	17
3J0L_g	-	-0.01	0.00	0.00	0	458	7	2	5	0	4
3J0L_a	-	0.55	0.50	0.62	8	1115	5	2	3	0	8
3J16_L	tRNA links to Rfam database...	0.35	0.33	0.38	10	2749	16	1	15	0	20
3RKF_A	Purine links to Rfam database...	0.73	0.59	0.91	20	2189	2	1	1	0	14
3SD1_A	THF links to Rfam database...	0.55	0.48	0.65	20	3885	11	2	9	0	22
3U4M_B	-	0.45	0.35	0.59	13	3138	9	1	8	0	24
3UZL_B	tRNA links to Rfam database...	0.73	0.62	0.85	23	3543	4	0	4	0	14
4A1C_2	5_8S_rRNA links to Rfam database...	0.13	0.15	0.12	5	11738	49	7	31	11	28
4A1C_3	5S_rRNA links to Rfam database...	0.73	0.61	0.87	33	7102	6	1	4	1	21
4AOB_A	SAM links to Rfam database...	0.50	0.40	0.63	17	4344	11	2	8	1	25
4ENB_A	crcB links to Rfam database...	0.73	0.58	0.92	11	1263	1	1	0	0	8
4ENC_A	crcB links to Rfam database...	0.45	0.42	0.50	8	1310	8	1	7	0	11

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