Figure 1.

The analysis pipeline. Solid black arrows represent the core analysis involving kinetic folding programs. For each non-coding RNA molecule examined, a reference sequence and known structural features were used to construct a multiple-sequence alignment (MSA). Non-redundant sequences were extracted from the MSA, and provided as input to three kinetic folding programs: Kinwalker, Kinéfold and RNAKinetics. The raw output, consisting of simulated folding pathways, was analyzed, and high-scoring helices were extracted. These predicted helices were mapped back onto the MSA. A comparison across programs then yielded helices with high scores from multiple programs. The dotted arrows represent an optional analysis involving the comparative helix prediction program Transat. A phylogenetic tree is required as input along with an MSA. FastTree2 is employed in our pipeline to estimate the phylogenetic tree based on the concatenated unpaired regions (83). The program detects and scores conserved helices, and outputs a ranked list of conserved helices with P-values. These are mapped onto the alignment along with the output of the three kinetic folding programs.