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 UNAFold - scored higher in this pairwise comparison

  4. Performance of Mastr(20) - scored lower in this pairwise comparison

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


Overview

Metric UNAFold Mastr(20)
MCC 0.662 > 0.506
Average MCC ± 95% Confidence Intervals 0.654 ± 0.145 > 0.489 ± 0.199
Sensitivity 0.663 > 0.337
Positive Predictive Value 0.667 < 0.769
Total TP 301 > 153
Total TN 51812 < 52064
Total FP 200 > 52
Total FP CONTRA 49 > 10
Total FP INCONS 101 > 36
Total FP COMP 50 > 6
Total FN 153 < 301
P-value 1.89649746203e-08

^top




Performance plots


  1. Comparison of performance of UNAFold and Mastr(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 UNAFold and Mastr(20)).

  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 UNAFold and Mastr(20)).

  3. Comparison of average Matthews Correlation Coefficients (MCCs) for UNAFold and Mastr(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 UNAFold and Mastr(20)).

^top





Performance of UNAFold - scored higher in this pairwise comparison

1. Total counts & total scores for UNAFold

Total Base Pair Counts
Total TP 301
Total TN 51812
Total FP 200
Total FP CONTRA 49
Total FP INCONS 101
Total FP COMP 50
Total FN 153
Total Scores
MCC 0.662
Average MCC ± 95% Confidence Intervals 0.654 ± 0.145
Sensitivity 0.663
Positive Predictive Value 0.667
Nr of predictions 15

^top



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

RNA Chain Rfam family MCC SENS PPV TP TN FP FP CONTRA FP INCONS FP COMP FN
3A2K_C 0.47 0.50 0.46 11 1084 13 3 10 0 11
3GX2_A 0.81 0.79 0.85 22 1423 5 2 2 1 6
3IVN_B 0.91 0.83 1.00 19 884 0 0 0 0 4
3IZ4_A 0.58 0.60 0.57 57 25436 53 12 31 10 38
3JYV_7 0.24 0.25 0.25 5 1091 16 7 8 1 15
3LA5_A 0.91 0.84 1.00 21 933 0 0 0 0 4
3NPB_A 0.85 0.78 0.94 29 2247 6 0 2 4 8
3O58_3 0.42 0.50 0.35 11 4733 34 5 15 14 11
3PDR_A 0.93 0.92 0.94 46 4791 5 1 2 2 4
3RKF_A 0.91 0.83 1.00 20 846 0 0 0 0 4
3SD1_A 0.77 0.76 0.79 22 1505 6 4 2 0 7
4A1C_2 0.19 0.25 0.15 5 4483 42 11 17 14 15
4AOB_A 0.60 0.59 0.63 17 1410 11 4 6 1 12
4ENB_A 0.85 0.73 1.00 11 461 2 0 0 2 4
4ENC_A 0.37 0.33 0.45 5 485 7 0 6 1 10

^top



Performance of Mastr(20) - scored lower in this pairwise comparison

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

Total Base Pair Counts
Total TP 153
Total TN 52064
Total FP 52
Total FP CONTRA 10
Total FP INCONS 36
Total FP COMP 6
Total FN 301
Total Scores
MCC 0.506
Average MCC ± 95% Confidence Intervals 0.489 ± 0.199
Sensitivity 0.337
Positive Predictive Value 0.769
Nr of predictions 15

^top



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

RNA Chain Rfam family MCC SENS PPV TP TN FP FP CONTRA FP INCONS FP COMP FN
3A2K_C 0.81 0.77 0.85 17 1088 3 0 3 0 5
3GX2_A 0.46 0.39 0.55 11 1429 10 2 7 1 17
3IVN_B 0.86 0.83 0.90 19 882 2 2 0 0 4
3IZ4_A 0.00 0.00 0.00 0 25536 0 0 0 0 95
3JYV_7 0.97 0.95 1.00 19 1092 2 0 0 2 1
3LA5_A 0.89 0.80 1.00 20 934 0 0 0 0 5
3NPB_A 0.34 0.30 0.41 11 2251 18 1 15 2 26
3O58_3 0.00 0.00 0.00 0 4764 0 0 0 0 22
3PDR_A 0.00 0.00 0.00 0 4840 0 0 0 0 50
3RKF_A 0.84 0.71 1.00 17 849 0 0 0 0 7
3SD1_A 0.73 0.72 0.75 21 1505 7 4 3 0 8
4A1C_2 0.00 0.00 0.00 0 4516 0 0 0 0 20
4AOB_A 0.42 0.34 0.53 10 1418 10 1 8 1 19
4ENB_A 0.44 0.20 1.00 3 469 0 0 0 0 12
4ENC_A 0.57 0.33 1.00 5 491 0 0 0 0 10

^top


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