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 RSpredict(20) - scored higher in this pairwise comparison

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

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


Overview

Metric RSpredict(20) NanoFolder
MCC 0.618 > 0.326
Average MCC ± 95% Confidence Intervals 0.569 ± 0.167 > 0.338 ± 0.143
Sensitivity 0.573 > 0.417
Positive Predictive Value 0.674 > 0.267
Total TP 184 > 134
Total TN 31204 > 30976
Total FP 129 < 436
Total FP CONTRA 31 < 131
Total FP INCONS 58 < 236
Total FP COMP 40 < 69
Total FN 137 < 187
P-value 2.18141491686e-08

^top




Performance plots


  1. Comparison of performance of RSpredict(20) and NanoFolder. 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 RSpredict(20) and NanoFolder).

  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 RSpredict(20) and NanoFolder).

  3. Comparison of average Matthews Correlation Coefficients (MCCs) for RSpredict(20) and NanoFolder. 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 RSpredict(20) and NanoFolder).

^top





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

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

Total Base Pair Counts
Total TP 184
Total TN 31204
Total FP 129
Total FP CONTRA 31
Total FP INCONS 58
Total FP COMP 40
Total FN 137
Total Scores
MCC 0.618
Average MCC ± 95% Confidence Intervals 0.569 ± 0.167
Sensitivity 0.573
Positive Predictive Value 0.674
Nr of predictions 13

^top



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

RNA Chain Rfam family MCC SENS PPV TP TN FP FP CONTRA FP INCONS FP COMP FN
3J20_1 0.67 0.50 0.91 10 1101 2 0 1 1 10
3J2L_3 0.71 0.59 0.87 20 2997 6 0 3 3 14
3J3D_C 0.82 0.84 0.80 16 948 4 3 1 0 3
3J3E_7 0.82 0.79 0.84 27 2709 7 0 5 2 7
3J3E_8 -0.01 0.00 0.00 0 2724 19 5 13 1 15
3J3F_8 0.42 0.47 0.38 9 4737 26 9 6 11 10
3J3F_7 0.94 0.94 0.94 34 2898 3 1 1 1 2
3ZEX_D 0.88 0.83 0.94 29 2765 7 0 2 5 6
3ZND_W 0.47 0.38 0.60 3 1186 9 0 2 7 5
4A1C_3 0.60 0.49 0.75 18 2739 8 0 6 2 19
4A1C_2 0.25 0.30 0.21 6 4487 30 12 11 7 14
4AOB_A 0.33 0.21 0.55 6 1426 5 1 4 0 23
4ENC_A 0.51 0.40 0.67 6 487 3 0 3 0 9

^top



Performance of NanoFolder - scored lower in this pairwise comparison

1. Total counts & total scores for NanoFolder

Total Base Pair Counts
Total TP 134
Total TN 30976
Total FP 436
Total FP CONTRA 131
Total FP INCONS 236
Total FP COMP 69
Total FN 187
Total Scores
MCC 0.326
Average MCC ± 95% Confidence Intervals 0.338 ± 0.143
Sensitivity 0.417
Positive Predictive Value 0.267
Nr of predictions 13

^top



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

RNA Chain Rfam family MCC SENS PPV TP TN FP FP CONTRA FP INCONS FP COMP FN
3J20_1 0.42 0.50 0.37 10 1085 20 5 12 3 10
3J2L_3 0.13 0.18 0.12 6 2969 46 14 31 1 28
3J3D_C 0.76 0.95 0.62 18 939 11 9 2 0 1
3J3E_7 0.44 0.53 0.38 18 2693 31 10 20 1 16
3J3E_8 -0.01 0.00 0.00 0 2707 47 11 24 12 15
3J3F_8 0.30 0.47 0.19 9 4714 57 21 17 19 10
3J3F_7 0.20 0.25 0.18 9 2883 43 10 32 1 27
3ZEX_D 0.30 0.37 0.27 13 2747 36 10 26 0 22
3ZND_W 0.24 0.38 0.16 3 1172 31 8 8 15 5
4A1C_3 0.58 0.68 0.51 25 2714 24 9 15 0 12
4A1C_2 -0.01 0.00 0.00 0 4469 61 18 29 14 20
4AOB_A 0.39 0.41 0.39 12 1406 20 4 15 1 17
4ENC_A 0.66 0.73 0.61 11 478 9 2 5 2 4

^top


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