# Transat—A Method for Detecting the Conserved Helices of Functional RNA Structures, Including Transient, Pseudo-Knotted and Alternative Structures

## Figure 16

Comparison of Transat (top figure) and RNAalifold P (bottom figure) for the *hok* data set.

In each figure, the x-axis represents the *hok* alignment. Each arc corresponds to a base-pair between the respective positions in the alignment. Arcs above the x-axis correspond to known base-pairs, whereas arcs below correspond to new base-pairs predicted by the respective program, i.e. they correspond to base-pairs that do not involve the same pair of nucleotide positions as any base-pair in the known structure(s). In the top figure, base-pairs predicted by Transat have non-black colours which indicate their reliability as estimated by Transat ( green, blue, orange) using a p-value threshold of . These base pairs can either be found above the x-axis, if they agree with a pair in the reference structure(s), or below, if they are new. In the bottom figure, base-pairs predicted by RNAalifold P have non-black colours which indicate their base-pairing probability in the Boltzmann ensemble of pseudo-knot free RNA secondary structures that we would expect in thermodynamic equilibrium ( green, blue, orange and red) using a pairing probability threshold of 5%. These base-pairs can either be found above the x-axis, if they agree with a pair in the reference structure(s), or below, if they are new. Transat predicts most helices of the known structures as well as three statistically significant conserved helices which may guide the structure formation, whereas RNAalifold P predicts only part of the known structures and contributes only a few novel base-pairs which extend a known helix by one or two base-pairs on either side, see also Figure 9.