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 TurboFold(20) - scored higher in this pairwise comparison

  4. Performance of Vsfold4 - scored lower in this pairwise comparison

  5. Compile and download dataset for TurboFold(20) & Vsfold4 [.zip] - may take several seconds...


Overview

Metric TurboFold(20) Vsfold4
MCC 0.796 > 0.619
Average MCC ± 95% Confidence Intervals 0.791 ± 0.111 > 0.662 ± 0.137
Sensitivity 0.762 > 0.603
Positive Predictive Value 0.838 > 0.647
Total TP 259 > 205
Total TN 22637 > 22629
Total FP 75 < 137
Total FP CONTRA 13 < 27
Total FP INCONS 37 < 85
Total FP COMP 25 = 25
Total FN 81 < 135
P-value 1.45096400029e-08

^top




Performance plots


  1. Comparison of performance of TurboFold(20) and Vsfold4. 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 TurboFold(20) and Vsfold4).

  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 TurboFold(20) and Vsfold4).

  3. Comparison of average Matthews Correlation Coefficients (MCCs) for TurboFold(20) and Vsfold4. 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 TurboFold(20) and Vsfold4).

^top





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

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

Total Base Pair Counts
Total TP 259
Total TN 22637
Total FP 75
Total FP CONTRA 13
Total FP INCONS 37
Total FP COMP 25
Total FN 81
Total Scores
MCC 0.796
Average MCC ± 95% Confidence Intervals 0.791 ± 0.111
Sensitivity 0.762
Positive Predictive Value 0.838
Nr of predictions 14

^top



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

RNA Chain Rfam family MCC SENS PPV TP TN FP FP CONTRA FP INCONS FP COMP FN
2L94_A 0.94 0.94 0.94 17 339 2 0 1 1 1
3AMU_B 0.81 0.79 0.83 15 1139 6 0 3 3 4
3J20_1 1.00 1.00 1.00 20 1092 1 0 0 1 0
3J20_0 0.74 0.76 0.73 16 1197 7 1 5 1 5
3J2L_3 0.92 0.91 0.94 31 2987 5 1 1 3 3
3RKF_A 0.91 0.83 1.00 20 846 0 0 0 0 4
3SD1_A 0.71 0.69 0.74 20 1506 7 2 5 0 9
3ZEX_D 0.86 0.80 0.93 28 2766 6 0 2 4 7
4A1C_3 0.86 0.84 0.89 31 2728 6 0 4 2 6
4A1C_2 0.22 0.25 0.21 5 4492 28 6 13 9 15
4AOB_A 0.67 0.59 0.77 17 1415 6 2 3 1 12
4ENB_A 0.77 0.60 1.00 9 463 0 0 0 0 6
4ENC_A 0.73 0.60 0.90 9 486 1 1 0 0 6
4FRG_B 0.93 0.88 1.00 21 1181 0 0 0 0 3

^top



Performance of Vsfold4 - scored lower in this pairwise comparison

1. Total counts & total scores for Vsfold4

Total Base Pair Counts
Total TP 205
Total TN 22629
Total FP 137
Total FP CONTRA 27
Total FP INCONS 85
Total FP COMP 25
Total FN 135
Total Scores
MCC 0.619
Average MCC ± 95% Confidence Intervals 0.662 ± 0.137
Sensitivity 0.603
Positive Predictive Value 0.647
Nr of predictions 14

^top



2. Individual counts for Vsfold4 [ download as .csv ]

RNA Chain Rfam family MCC SENS PPV TP TN FP FP CONTRA FP INCONS FP COMP FN
2L94_A 0.94 0.89 1.00 16 341 1 0 0 1 2
3AMU_B 0.73 0.74 0.74 14 1138 8 0 5 3 5
3J20_1 0.75 0.75 0.75 15 1092 7 0 5 2 5
3J20_0 0.95 0.95 0.95 20 1198 2 1 0 1 1
3J2L_3 0.71 0.71 0.73 24 2987 12 3 6 3 10
3RKF_A 0.91 0.83 1.00 20 846 0 0 0 0 4
3SD1_A 0.23 0.24 0.24 7 1504 22 5 17 0 22
3ZEX_D 0.66 0.63 0.71 22 2765 10 4 5 1 13
4A1C_3 0.29 0.30 0.31 11 2727 26 4 21 1 26
4A1C_2 0.36 0.40 0.33 8 4492 28 4 12 12 12
4AOB_A 0.50 0.48 0.54 14 1411 13 3 9 1 15
4ENB_A 0.85 0.73 1.00 11 461 0 0 0 0 4
4ENC_A 0.77 0.60 1.00 9 487 0 0 0 0 6
4FRG_B 0.60 0.58 0.64 14 1180 8 3 5 0 10

^top


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