CompaRNA - on-line benchmarks of RNA structure prediction methods
Home

Methods
Datasets
Rankings
RNA 2D Atlas

Help
FAQ

Contact us
RSS feeds
Twitter

Table of contents:

  1. Overview

  2. Performance Plots

  3. Performance of Carnac(20) - scored higher in this pairwise comparison

  4. Performance of Vsfold5 - scored lower in this pairwise comparison

  5. Compile and download dataset for Carnac(20) & Vsfold5 [.zip] - may take several seconds...


Overview

Metric Carnac(20) Vsfold5
MCC 0.662 > 0.495
Average MCC ± 95% Confidence Intervals 0.610 ± 0.164 > 0.538 ± 0.201
Sensitivity 0.469 < 0.480
Positive Predictive Value 0.938 > 0.519
Total TP 213 < 218
Total TN 52036 > 51843
Total FP 24 < 242
Total FP CONTRA 6 < 52
Total FP INCONS 8 < 150
Total FP COMP 10 < 40
Total FN 241 > 236
P-value 2.84041289224e-08

^top




Performance plots


  1. Comparison of performance of Carnac(20) and Vsfold5. 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 Carnac(20) and Vsfold5).

  2. 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 Carnac(20) and Vsfold5).

  3. Comparison of average Matthews Correlation Coefficients (MCCs) for Carnac(20) and Vsfold5. 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 Carnac(20) and Vsfold5).

^top





Performance of Carnac(20) - scored higher in this pairwise comparison

1. Total counts & total scores for Carnac(20)

Total Base Pair Counts
Total TP 213
Total TN 52036
Total FP 24
Total FP CONTRA 6
Total FP INCONS 8
Total FP COMP 10
Total FN 241
Total Scores
MCC 0.662
Average MCC ± 95% Confidence Intervals 0.610 ± 0.164
Sensitivity 0.469
Positive Predictive Value 0.938
Nr of predictions 15

^top



2. Individual counts for Carnac(20) [ download as .csv ]

RNA Chain Rfam family MCC SENS PPV TP TN FP FP CONTRA FP INCONS FP COMP FN
3A2K_C 0.88 0.77 1.00 17 1091 0 0 0 0 5
3GX2_A 0.68 0.46 1.00 13 1436 0 0 0 0 15
3IVN_B 0.91 0.83 1.00 19 884 0 0 0 0 4
3IZ4_A 0.52 0.31 0.88 29 25503 4 3 1 0 66
3JYV_7 0.81 0.70 0.93 14 1096 3 0 1 2 6
3LA5_A 0.91 0.84 1.00 21 933 0 0 0 0 4
3NPB_A 0.46 0.22 1.00 8 2270 3 0 0 3 29
3O58_3 0.60 0.36 1.00 8 4756 1 0 0 1 14
3PDR_A 0.75 0.58 0.97 29 4810 3 1 0 2 21
3RKF_A 0.91 0.83 1.00 20 846 0 0 0 0 4
3SD1_A 0.72 0.55 0.94 16 1516 1 0 1 0 13
4A1C_2 0.42 0.25 0.71 5 4509 3 0 2 1 15
4AOB_A 0.59 0.48 0.74 14 1418 6 2 3 1 15
4ENB_A 0.00 0.00 0.00 0 472 0 0 0 0 15
4ENC_A 0.00 0.00 0.00 0 496 0 0 0 0 15

^top



Performance of Vsfold5 - scored lower in this pairwise comparison

1. Total counts & total scores for Vsfold5

Total Base Pair Counts
Total TP 218
Total TN 51843
Total FP 242
Total FP CONTRA 52
Total FP INCONS 150
Total FP COMP 40
Total FN 236
Total Scores
MCC 0.495
Average MCC ± 95% Confidence Intervals 0.538 ± 0.201
Sensitivity 0.480
Positive Predictive Value 0.519
Nr of predictions 15

^top



2. Individual counts for Vsfold5 [ download as .csv ]

RNA Chain Rfam family MCC SENS PPV TP TN FP FP CONTRA FP INCONS FP COMP FN
3A2K_C 0.95 0.91 1.00 20 1088 0 0 0 0 2
3GX2_A 0.61 0.57 0.67 16 1425 9 0 8 1 12
3IVN_B 0.88 0.78 1.00 18 885 0 0 0 0 5
3IZ4_A 0.31 0.31 0.33 29 25448 65 14 45 6 66
3JYV_7 -0.02 0.00 0.00 0 1092 19 4 15 0 20
3LA5_A -0.02 0.00 0.00 0 938 16 0 16 0 25
3NPB_A 0.62 0.54 0.71 20 2250 10 1 7 2 17
3O58_3 0.51 0.59 0.45 13 4735 28 11 5 12 9
3PDR_A 0.69 0.64 0.74 32 4797 13 3 8 2 18
3RKF_A 0.90 0.92 0.88 22 841 3 3 0 0 2
3SD1_A 0.13 0.14 0.15 4 1507 22 5 17 0 25
4A1C_2 0.33 0.40 0.28 8 4487 37 8 13 16 12
4AOB_A 0.21 0.21 0.25 6 1413 19 2 16 1 23
4ENB_A 1.00 1.00 1.00 15 457 0 0 0 0 0
4ENC_A 0.97 1.00 0.94 15 480 1 1 0 0 0

^top


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