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 Afold - scored higher in this pairwise comparison

  4. Performance of RSpredict(seed) - scored lower in this pairwise comparison

  5. Compile and download dataset for Afold & RSpredict(seed) [.zip] - may take several seconds...


Overview

Metric Afold RSpredict(seed)
MCC 0.551 > 0.351
Average MCC ± 95% Confidence Intervals 0.617 ± 0.185 > 0.311 ± 0.188
Sensitivity 0.597 > 0.205
Positive Predictive Value 0.514 < 0.608
Total TP 253 > 87
Total TN 67468 < 67817
Total FP 296 > 66
Total FP CONTRA 100 > 17
Total FP INCONS 139 > 39
Total FP COMP 57 > 10
Total FN 171 < 337
P-value 2.12160100812e-08

^top




Performance plots


  1. Comparison of performance of Afold and RSpredict(seed). 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 Afold and RSpredict(seed)).

  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 Afold and RSpredict(seed)).

  3. Comparison of average Matthews Correlation Coefficients (MCCs) for Afold and RSpredict(seed). 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 Afold and RSpredict(seed)).

^top





Performance of Afold - scored higher in this pairwise comparison

1. Total counts & total scores for Afold

Total Base Pair Counts
Total TP 253
Total TN 67468
Total FP 296
Total FP CONTRA 100
Total FP INCONS 139
Total FP COMP 57
Total FN 171
Total Scores
MCC 0.551
Average MCC ± 95% Confidence Intervals 0.617 ± 0.185
Sensitivity 0.597
Positive Predictive Value 0.514
Nr of predictions 14

^top



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

RNA Chain Rfam family MCC SENS PPV TP TN FP FP CONTRA FP INCONS FP COMP FN
2LC8_A 0.64 0.61 0.69 11 512 6 0 5 1 7
2ZZN_D 0.91 0.91 0.91 20 962 3 2 0 1 2
3A2K_C 0.46 0.50 0.44 11 1083 14 3 11 0 11
3A3A_A 0.93 0.87 1.00 26 1474 0 0 0 0 4
3IVN_B 0.91 0.83 1.00 19 884 0 0 0 0 4
3IYQ_A 0.23 0.33 0.17 17 22339 97 47 37 13 34
3IZ4_A 0.52 0.57 0.48 54 25423 63 27 32 4 41
3JYV_7 -0.02 0.00 0.00 0 1088 23 4 19 0 20
3LA5_A 0.91 0.84 1.00 21 933 0 0 0 0 4
3NPB_A 0.82 0.73 0.93 27 2249 6 0 2 4 10
3O58_3 0.41 0.50 0.34 11 4732 38 6 15 17 11
3RKF_A 0.91 0.83 1.00 20 846 0 0 0 0 4
4A1C_2 0.19 0.25 0.15 5 4483 43 11 17 15 15
4ENB_A 0.81 0.73 0.92 11 460 3 0 1 2 4

^top



Performance of RSpredict(seed) - scored lower in this pairwise comparison

1. Total counts & total scores for RSpredict(seed)

Total Base Pair Counts
Total TP 87
Total TN 67817
Total FP 66
Total FP CONTRA 17
Total FP INCONS 39
Total FP COMP 10
Total FN 337
Total Scores
MCC 0.351
Average MCC ± 95% Confidence Intervals 0.311 ± 0.188
Sensitivity 0.205
Positive Predictive Value 0.608
Nr of predictions 14

^top



2. Individual counts for RSpredict(seed) [ download as .csv ]

RNA Chain Rfam family MCC SENS PPV TP TN FP FP CONTRA FP INCONS FP COMP FN
2LC8_A -0.03 0.00 0.00 0 515 13 0 13 0 18
2ZZN_D 0.38 0.18 0.80 4 979 1 1 0 0 18
3A2K_C -0.01 0.00 0.00 0 1106 2 0 2 0 22
3A3A_A 0.53 0.37 0.79 11 1486 3 0 3 0 19
3IVN_B 0.81 0.70 0.94 16 886 1 1 0 0 7
3IYQ_A 0.16 0.10 0.26 5 22421 21 10 4 7 46
3IZ4_A 0.26 0.11 0.63 10 25520 7 2 4 1 85
3JYV_7 -0.01 0.00 0.00 0 1108 4 0 3 1 20
3LA5_A 0.82 0.68 1.00 17 937 0 0 0 0 8
3NPB_A -0.01 0.00 0.00 0 2274 4 0 4 0 37
3O58_3 0.28 0.14 0.60 3 4759 2 0 2 0 19
3RKF_A 0.84 0.75 0.95 18 847 1 1 0 0 6
4A1C_2 0.00 0.00 0.00 0 4512 5 2 2 1 20
4ENB_A 0.34 0.20 0.60 3 467 2 0 2 0 12

^top


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