PMC

The two canonical forms of the IRE.

Schematic diagram of the structure of the algorithm.

Highest scoring motif occurrences output by RNAProfile on the Drosophila nanos dataset.

Highest scoring motif occurrences in the GPX4 sequences combined with the selenoprotein M dataset.

Secondary structure models of SECIS stem–loop elements: type I (left) and type II (right).

Highest scoring motif occurrences output by RNAProfile on the GPX4 dataset. The last region, with a significantly low fitness value, comes from a non-selenoprotein 3′-UTR.

Highest scoring motif occurrences output by RNAProfile on the RNAse P RNA dataset, corresponding to consensus helices 8 and 9 (see also Figure ).

Highest scoring motif occurrences output by RNAProfile on the IRE dataset with their respective energy and fitness value. Note that the last three regions (reported in pseudogenes) have a much lower fitness value, thus very unlikely to be real IRE instances.

Secondary structure of human RNase P RNA [structure adapted from ()]. Numbered helices form the consensus secondary structure of the molecule, conserved (despite no sequence conservation) in bacteria throughout eukaryotes.

The results on the atypical IRE dataset. The first four instances, corresponding to the IREs, were included in the highest-scoring profile of four regions (and also the best two and three regions groups contained IRE instances). In the following two iterations, the best profile still contained the same four regions, plus the two shown at the bottom of the figure, unlikely to be IRE instances given the low fitness value.