Skip to main content
Fig. 3 | Genome Biology

Fig. 3

From: DotAligner: identification and clustering of RNA structure motifs

Fig. 3

Classification of known RNA structures. a Receiving operator characteristic (ROC) curves measuring the classification accuracy of the surveyed algorithms by contrasting their computed similarity matrices to a binary classification matrix of Rfam sequences (1 if the sequences are in the same family or 0 if different). High PID = 56–95 % pairwise sequence identity from the provided Rfam alignment; low PID = 1–55 %. b Precision vs recall curve. c Area under the curve (AUC) of ROC values with 95 % confidence intervals for the top four performing algorithms across five ranges of pairwise sequence identity, as calculated from a variant of the Needleman–Wunsch algorithm with free end gaps. The three replicates correspond to stochastically sampled sequences from Rfam 12.3 (see Additional file 2: Table S1). d Runtime distribution of single-thread computation on a 2.6 GHz AMD Opteron processor (note, a fixed upper limit of 120 s was imposed for CARNA). AUC area under the curve

Back to article page