Illustration of the evaluation of the prediction and the performance of the seven programs

A. Illustration of the evaluation of prediction. The reference (left panel) and predicted (right panel) structures of a 6-nt RNA sequence are represented by arc diagrams. In this prediction, the TP is base pair 1–6 (green in right panel) as it is present in both reference and predicted structures; the FN is base pair 2–5 as it is present only in the reference structure; the FP is base pair 2–4 (red in right panel) as it is present only in the predicted structure; and the TNs are base pairs 1–2, 1–3, 1–4, 1–5, 2–3, 2–6, 3–4, 3–5, 3–6, 4–5, 4–6, and 5–6 as they are absent in both reference and predicted structures (i.e., all the possible base pairs i–j, i, j ∈ [1,6], i < j, with the exclusion of the TP, FN, and FP). Thus, for this prediction, the precision is 0.5, the sensitivity is 0.5, the F1 score is 0.5, and the MCC is ∼ 0.42. B. Box plots showing the precision, sensitivity, F1 score, and MCC across the 32 ribozyme sequences for each program. C. Heatmap showing the MCCs of the predictions obtained by each program for each ribozyme. The numbers on the left side of the heatmap are the median MCCs for each program. TP, true positive; FN, false negative; FP, false positive; TN, true negative; MCC, Matthews correlation coefficient; HHR, hammerhead; HDV, hepatitis delta virus; TW, twister; TS, twister sister; VS, Varkud satellite.