Skip to main content
. 2015 Nov 3;16(11):26303–26317. doi: 10.3390/ijms161125952

Table 4.

Performance of the state-of-the-art methods for RNA-binding site prediction.

Methods Data Set Performance Reference Feature
ACC SEN SPE AUC MCC Strength F-Measure Precision
PiRaNhA RB75 - - - 0.822 0.435 - - - [8] Sequence-based
PPRInt RB75 - - - 0.779 0.339 - - - [8]
RB172 0.71 - 0.25 0.66 - - [28]
RB344 0.70 0.45 0.82 0.68 0.28 - 0.49 0.53 [26]
BindN RB75 - - - 0.733 0.297 - - - [8]
RB172 0.75 - - - 0.23 0.64 - - [28]
BindN+ RB75 - - - 0.821 0.397 - - - [8]
RB172 0.79 - - - 0.34 0.71 - - [28]
RB344 0.72 0.32 0.89 0.68 0.26 - 0.41 0.56 [26]
RNABindR RB75 - - - 0.708 0.317 - - - [8]
RNABindR v2.0 RB172 0.66 - - - 0.27 0.69 - - [28]
PRBR RB75 - - - N/A a 0.294 - - - [8]
NAPS RB75 - - - 0.679 0.215 - - - [8]
RB172 0.66 - - - 0.17 0.61 - - [28]
RNAProB RB172 0.82 - - - 0.22 0.60 - - [28]
KYG * RB75 - - - N/A 0.382 - - - [8] Structure-based
DRNA * RB75 - - - N/A 0.382 - - - [8]
RB344 0.75 0.21 0.94 N/A 0.22 - 0.31 0.54 [26]
OPRA * RB75 - - - N/A 0.296 - - - [8]
Ren’s method RB344 0.68 0.48 0.76 0.68 0.26 - 0.48 0.48 [26,83]
Meta-predictor b RB75 - - - 0.835 0.460 - - - [8,34]

a N/A—not available; MCC—Matthews Correlation Coefficient; AUC—area under curve; SEN—sensitivity; SPE—specificity; b Meta-predictor developed based on top three sequence-based methods according to authors benchmark (PiRaNhA, PPRInt and BindN+); * The meta-predictor is composed of those methods labeled with asterisk.