Table 1. Correspondence between phylogenetically-determined single-stranded regions and peaks on the probability profile and improvement in predictions over minimum free energy structure.

RNA sequence	Accession no.	Length (nt)	PC (%)a	PC1 (%)b	PC2 (%)c	PC3 (%)d	PI (%)e
Escherichia coli tRNAAla	X66515	76	100	100	100	0	20
Xenopus laevis oocyte 5S rRNA	K02695	120	100	100	100	25	28
Escherichia coli 16S rRNA domain II	J01695	353	82	100	50	33	29
Escherichia coli RNase P	V00338	377	100	100	58	50	40
Tetrahymena thermophila LSU group I intron	V01416	413	95	88	67	29	19

^aP_C is the percentage of phylogenetically-determined single-stranded regions (region here is either a sequence of four consecutive nucleotides or several such sequences in a row) that correspond to peaks (regardless of the magnitude of the maximum probability) in the probability profile in Figure 2.

^bFor peaks with a maximum probability ≥0.5, P_C1 is the percentage of these peaks that correspond to single-stranded regions.

^cP_C2 is the percentage of the correspondence for peaks with a maximum probability between 0.2 and 0.5.

^dP_C3 is the percentage of the correspondence for peaks with maximum probability <0.2.

^eA probability profile predicts more single-stranded regions in the phylogenetic structure than the minimum free energy structure (Figs 1B and 2). P_I is the percentage of improvement in the prediction by the probability profile over the MFE structure. This is computed by the number of regions missed by the MFE structure but predicted by the probability profile divided by the total number of single-stranded regions in the phylogenetic structure (e.g., seven for Xlo 5S RNA) and multiplied by 100%.