PMC

Illustration of a potential mechanism for the selection of exon 17. When the stem from the blue sequences is implemented, it potentially obstructs the splicing branch point of intron 16-17.1 inhibiting the selection of exon 17.1 and resulting in the selection of exon 17.2. If the competing stem from the red sequences is implemented, then, under the additional assumption that a pseudoknot involving both interactions is not implemented, exon 16 is allowed to be spliced directly into exon 17.1.

Multiple alignment of the binding sites of orthologous introns showing conservation and drift. (a) The sequence 5'-GGCAGUUUUCAA-3' upstream of exon 6.25 of Drosophila melanogaster is perfectly conserved in the orthologous exons (Table 1) across all species. (b) The binding sequence upstream of exon 6.48 of D. melanogaster drifts into different binding sequences in the orthologous exons of other species. Areas indicated in red are the predicted binding sites.

The structure of the Dscam gene in six Drosophila spp. The mature mRNA of each gene contains 22 exons, four of which (exon 4, exon 6, exon 9 and exon 17, shown in green, blue, yellow and red, respectively) are chosen from a cluster of alternative exons. The remaining 18 exons, shown in metallic gray, are always selected. Introns are not shown. Red lines indicate an example choice of exons.

Output of the genetic algorithm showing convergence to anchor sequence. Starting from a completely random sequence, the genetic algorithm is able to converge to a sequence that is a substring of the full anchor sequence. (a) The fitness progression of one of the runs of the genetic algorithm. The genetic algorithm searches in the space of all possible sequences to converge to a high fitness sequence that matches the anchor sequence. (b) The evolution of the estimated anchor sequence at each iteration of the genetic algorithm. The output converges to a substring of the actual anchor sequence, namely AAAUUGAAAACUGCCUGAAUGUUGGGAUAGGGUACUCGACAA.

Dot plots illustrating the binding sites upstream of each alternative exon 6 in Drosophila melanogaster. The horizontal axis shows the distance in nucleotides from the exon to be chosen. The vertical axis indicates the nucleotides of the anchor sequence. The dot plots were created using the '8 out of 11' method and were further processed to filter out any dots that are not part of a continuous line of length at least 4. Line lengths should be seen as approximate indicators and not as precise measures of binding strength.

Illustration of variable window binding using sequence data from Drosophila melanogaster. The anchor sequence (conserved in all six species) is located downstream from exon 5 and is shown as the 'lower' strand, whereas its binding site is located upstream of the exon to be selected and is shown as the 'upper' strand. There are multiple alternative mutually exclusive pre-mRNA local secondary structures, three of which are shown. The figure is not drawn to scale. (a) Competition among multiple regions for pairing with the anchor sequence provides mutual exclusivity. (b) The binding site when exon 6.25 is selected to be connected with exon 5 corresponds to a window close to one end of the anchor sequence. (c) The binding site when exon 6.48 is selected to be connected with exon 5 corresponds to a window in the middle of the anchor sequence. (d) The binding site when exon 6.3 is selected to be connected with exon 5 corresponds to a window close to the other end of the anchor sequence.