Figure 1
From: Sampled ensemble neutrality as a feature to classify potential structured RNAs
SEN calculated neutrality has larger separation between structured and unstructured sequence. Distribution of neutrality values from Dataset2 compare the performance of various distance functions (A) RNAmute dot-bracket representation, (B) RNAmute Shapiro representation, (C) bp-distance, (D) Pearson’s Correlation Coefficient (PCC), and (E) Sampled Ensemble Neutrality (SEN). The 3’ and 5’ flanking region used for negatives are referred to as 3prime and 5prime, respectively. The SEN on the positive test set has a larger separation between the negatives, compared to other measures. All distance metrics show unstructured sequence to be low on their respective scales. Lastly, the SEN uses a large dynamic range of values compared to the base pair distance metric which will increase its sensitivity between highly similar structures.