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 PETfold_pre2.0(seed) - scored higher in this pairwise comparison

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

  5. Compile and download dataset for PETfold_pre2.0(seed) & MaxExpect [.zip] - may take several seconds...


Overview

Metric PETfold_pre2.0(seed) MaxExpect
MCC 0.866 > 0.617
Average MCC ± 95% Confidence Intervals 0.837 ± 0.059 > 0.584 ± 0.118
Sensitivity 0.841 > 0.627
Positive Predictive Value 0.892 > 0.609
Total TP 858 > 640
Total TN 473383 > 473294
Total FP 280 < 567
Total FP CONTRA 35 < 94
Total FP INCONS 69 < 317
Total FP COMP 176 > 156
Total FN 162 < 380
P-value 5.06544643719e-08

^top




Performance plots


  1. Comparison of performance of PETfold_pre2.0(seed) and MaxExpect. 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 PETfold_pre2.0(seed) and MaxExpect).

  2. Comparison of performance of PETfold_pre2.0(seed) and MaxExpect. 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 PETfold_pre2.0(seed) and MaxExpect).

  3. 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 PETfold_pre2.0(seed) and MaxExpect).

  4. 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 PETfold_pre2.0(seed) and MaxExpect).

  5. Comparison of average Matthews Correlation Coefficients (MCCs) for PETfold_pre2.0(seed) and MaxExpect. 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 PETfold_pre2.0(seed) and MaxExpect).

  6. Comparison of average Matthews Correlation Coefficients (MCCs) for PETfold_pre2.0(seed) and MaxExpect. 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 PETfold_pre2.0(seed) and MaxExpect).

^top





Performance of PETfold_pre2.0(seed) - scored higher in this pairwise comparison

1. Total counts & total scores for PETfold_pre2.0(seed)

Total Base Pair Counts
Total TP 858
Total TN 473383
Total FP 280
Total FP CONTRA 35
Total FP INCONS 69
Total FP COMP 176
Total FN 162
Total Scores
MCC 0.866
Average MCC ± 95% Confidence Intervals 0.837 ± 0.059
Sensitivity 0.841
Positive Predictive Value 0.892
Nr of predictions 26

^top



2. Individual counts for PETfold_pre2.0(seed) [ download as .csv ]

RNA Chain Rfam family MCC SENS PPV TP TN FP FP CONTRA FP INCONS FP COMP FN
2LC8_A 0.41 0.33 0.55 6 517 5 0 5 0 12
3J16_L 1.00 1.00 1.00 21 1138 1 0 0 1 0
3J20_1 1.00 1.00 1.00 20 1092 2 0 0 2 0
3J20_2 0.86 0.86 0.86 356 421956 129 18 38 73 56
3J20_0 0.95 0.95 0.95 20 1198 2 1 0 1 1
3J2L_3 0.97 0.94 1.00 32 2988 5 0 0 5 2
3J3D_C 0.90 0.95 0.86 18 947 4 3 0 1 1
3J3E_8 0.71 0.67 0.77 10 2729 8 1 2 5 5
3J3E_7 0.97 0.97 0.97 33 2707 6 0 1 5 1
3J3F_8 0.86 0.84 0.89 16 4743 13 2 0 11 3
3J3F_7 0.99 0.97 1.00 35 2899 4 0 0 4 1
3J3V_B 0.90 0.89 0.92 24 2630 14 0 2 12 3
3UZL_B 0.93 0.88 1.00 14 1279 8 0 0 8 2
3W1K_J 0.85 0.81 0.89 25 1650 4 2 1 1 6
3W3S_B 0.80 0.73 0.89 24 1962 6 1 2 3 9
3ZEX_D 0.96 0.94 0.97 33 2762 6 0 1 5 2
3ZEX_C 0.70 0.59 0.85 17 5354 12 1 2 9 12
3ZND_W 0.67 0.75 0.60 6 1181 16 0 4 12 2
4A1C_3 1.00 1.00 1.00 37 2726 2 0 0 2 0
4A1C_2 0.79 0.75 0.83 15 4498 12 1 2 9 5
4AOB_A 0.85 0.79 0.92 23 1412 4 0 2 2 6
4ENB_A 0.61 0.53 0.73 8 461 5 1 2 2 7
4ENC_A 0.61 0.53 0.73 8 485 5 1 2 2 7
4FRG_B 0.87 0.83 0.91 20 1180 3 0 2 1 4
4FRN_A 0.83 0.79 0.88 22 1823 3 2 1 0 6
4JF2_A 0.76 0.63 0.94 15 1066 1 1 0 0 9

^top



Performance of MaxExpect - scored lower in this pairwise comparison

1. Total counts & total scores for MaxExpect

Total Base Pair Counts
Total TP 640
Total TN 473294
Total FP 567
Total FP CONTRA 94
Total FP INCONS 317
Total FP COMP 156
Total FN 380
Total Scores
MCC 0.617
Average MCC ± 95% Confidence Intervals 0.584 ± 0.118
Sensitivity 0.627
Positive Predictive Value 0.609
Nr of predictions 26

^top



2. Individual counts for MaxExpect [ 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 0 513 15 3 12 0 18
3J16_L 0.34 0.33 0.37 7 1140 12 0 12 0 14
3J20_1 0.75 0.75 0.75 15 1092 6 0 5 1 5
3J20_2 0.60 0.63 0.58 258 421926 239 28 156 55 154
3J20_0 0.53 0.57 0.50 12 1195 13 2 10 1 9
3J2L_3 0.80 0.82 0.78 28 2984 12 1 7 4 6
3J3D_C 0.47 0.53 0.43 10 945 13 5 8 0 9
3J3E_8 0.10 0.13 0.09 2 2719 32 5 16 11 13
3J3E_7 0.60 0.59 0.63 20 2709 14 2 10 2 14
3J3F_8 0.34 0.42 0.29 8 4733 39 10 10 19 11
3J3F_7 0.94 0.94 0.94 34 2898 4 1 1 2 2
3J3V_B 0.71 0.74 0.69 20 2627 17 2 7 8 7
3UZL_B 1.00 1.00 1.00 16 1277 8 0 0 8 0
3W1K_J 0.97 0.97 0.97 30 1647 1 1 0 0 1
3W3S_B 0.97 0.94 1.00 31 1958 1 0 0 1 2
3ZEX_D 0.82 0.77 0.87 27 2765 11 0 4 7 8
3ZEX_C 0.54 0.45 0.65 13 5354 17 2 5 10 16
3ZND_W 0.24 0.38 0.16 3 1172 25 9 7 9 5
4A1C_3 0.83 0.81 0.86 30 2728 7 0 5 2 7
4A1C_2 0.19 0.25 0.16 5 4484 41 10 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 0 0 0 0 4
4ENC_A 0.36 0.33 0.42 5 484 8 0 7 1 10
4FRG_B 0.29 0.25 0.35 6 1185 11 4 7 0 18
4FRN_A 0.51 0.46 0.57 13 1825 10 5 5 0 15
4JF2_A 0.89 0.79 1.00 19 1063 0 0 0 0 5

^top


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