Fig 1.
Base pair probability matrix.
Table 1.
Comparison of the results obtained by the MFE RNAfold (Column 3), and RNAstructure when it uses MFE (RNAstr. MFE, Column 4) and real SHAPE data (RNAstr. real, Column 5), and our proposed method (ESD-Fold) when the real and simulated SHAPE data are employed (real and simulated, Columns 6-7).
The computational time (minutes) of ESD-Fold is represented in Column 8. Here, the values are rounded to one decimal digit.
Table 2.
The average values of TP, FP, FN, Sn, PPV, Accuracy ± standard deviation (Acc ± sd), and number of generations for 100 independent executions of the proposed method over the fifteen RNA sequences.
Here, the values are rounded to two decimal digits.
Fig 2.
Boxplots of simulated SHAPE data prediction accuracy versus accuracy of the MFE prediction of RNAstructure software.
In each box, the midline marks the median accuracy for 100 predictions.
Fig 3.
Boxplots of simulated SHAPE data prediction accuracy versus accuracy of the real SHAPE prediction of RNAstructure software.
In each box, the midline marks the median accuracy for 100 predictions.
Fig 4.
Boxplots of simulated SHAPE data prediction accuracy versus accuracy of MEA structure of RNAfold software.
In each box, the midline marks the median accuracy for 100 predictions.
Table 3.
PPV(M) shows the number of native base pairs in the MFE structure M.
PPVs corresponding to the sets of common base pairs and remaining MFE base pairs were computed. Values in columns 3 and 4 are the average of 100 independent executions of ESD-Fold ± standard deviation. Here, the values are rounded to two decimal digits.
Table 4.
Comparison of the results obtained by the MFE RNAfold (Column 3), and RNAstructure when it uses MFE (Column 4), and our proposed method (ESD-Fold) when the simulated SHAPE data is employed (Column 5).
The average of computational time (minutes) of ESD-Fold on each family is shown in Column 6. Here, the values are rounded to one decimal digit.
Table 5.
The average values of TP, FP, FN, Sn, PPV, Accuracy ± standard deviation (Acc ± sd), and number of generations for 200 independent executions of the proposed method over each 8 families of the RNA-STRAND dataset.
Here, the values are rounded to two decimal digits.