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 Multilign(seed) - scored higher in this pairwise comparison

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

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


Overview

Metric Multilign(seed) UNAFold
MCC 0.485 > 0.469
Average MCC ± 95% Confidence Intervals 0.399 ± 0.173 < 0.466 ± 0.161
Sensitivity 0.386 < 0.437
Positive Predictive Value 0.619 > 0.514
Total TP 83 < 94
Total TN 22737 > 22688
Total FP 52 < 101
Total FP CONTRA 6 < 14
Total FP INCONS 45 < 75
Total FP COMP 1 < 12
Total FN 132 > 121
P-value 5.59774301112e-05

^top




Performance plots


  1. Comparison of performance of Multilign(seed) and UNAFold. 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 Multilign(seed) and UNAFold).

  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 Multilign(seed) and UNAFold).

  3. Comparison of average Matthews Correlation Coefficients (MCCs) for Multilign(seed) and UNAFold. 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 Multilign(seed) and UNAFold).

^top





Performance of Multilign(seed) - scored higher in this pairwise comparison

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

Total Base Pair Counts
Total TP 83
Total TN 22737
Total FP 52
Total FP CONTRA 6
Total FP INCONS 45
Total FP COMP 1
Total FN 132
Total Scores
MCC 0.485
Average MCC ± 95% Confidence Intervals 0.399 ± 0.173
Sensitivity 0.386
Positive Predictive Value 0.619
Nr of predictions 13

^top



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

RNA Chain Rfam family MCC SENS PPV TP TN FP FP CONTRA FP INCONS FP COMP FN
RFA_00416 0.57 0.53 0.62 8 1472 6 1 4 1 7
RFA_00654 0.00 0.00 0.00 0 2414 1 0 1 0 18
RFA_00658 0.00 0.00 0.00 0 1128 0 0 0 0 14
RFA_00664 0.00 0.00 0.00 0 990 0 0 0 0 14
RFA_00708 0.00 0.00 0.00 0 1035 0 0 0 0 14
RFA_00767 0.63 0.56 0.71 10 1877 4 0 4 0 8
RFA_00768 0.61 0.56 0.67 10 1876 5 0 5 0 8
RFA_00769 0.55 0.56 0.56 10 1935 8 3 5 0 8
RFA_00770 0.68 0.56 0.83 10 2004 2 0 2 0 8
RFA_00773 0.59 0.56 0.63 10 1937 6 1 5 0 8
RFA_00779 0.61 0.56 0.67 10 1938 5 0 5 0 8
RFA_00808 0.60 0.56 0.64 9 2002 5 0 5 0 7
RFA_00809 0.37 0.38 0.38 6 2129 10 1 9 0 10

^top



Performance of UNAFold - scored lower in this pairwise comparison

1. Total counts & total scores for UNAFold

Total Base Pair Counts
Total TP 94
Total TN 22688
Total FP 101
Total FP CONTRA 14
Total FP INCONS 75
Total FP COMP 12
Total FN 121
Total Scores
MCC 0.469
Average MCC ± 95% Confidence Intervals 0.466 ± 0.161
Sensitivity 0.437
Positive Predictive Value 0.514
Nr of predictions 13

^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
RFA_00416 0.93 0.93 0.93 14 1470 4 0 1 3 1
RFA_00654 0.27 0.28 0.28 5 2397 13 2 11 0 13
RFA_00658 0.59 0.50 0.70 7 1118 5 0 3 2 7
RFA_00664 -0.01 0.00 0.00 0 980 11 0 10 1 14
RFA_00708 0.40 0.29 0.57 4 1028 3 0 3 0 10
RFA_00767 0.63 0.56 0.71 10 1877 4 0 4 0 8
RFA_00768 0.48 0.44 0.53 8 1876 7 1 6 0 10
RFA_00769 0.54 0.56 0.53 10 1934 9 4 5 0 8
RFA_00770 0.61 0.56 0.67 10 2001 8 0 5 3 8
RFA_00773 0.71 0.56 0.91 10 1942 4 1 0 3 8
RFA_00779 0.55 0.56 0.56 10 1935 8 3 5 0 8
RFA_00808 -0.01 0.00 0.00 0 2001 15 2 13 0 16
RFA_00809 0.37 0.38 0.38 6 2129 10 1 9 0 10

^top


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