Pathway prediction by bioinformatic analysis of the untranslated regions of the CFTR mRNA.

Mining the information contained within the genetic code in untranslated regions has proven difficult because of the ambiguity of microRNA and protein binding sites. This manuscript describes a bioinformatic screen that identifies long sequences with partial identity to the untranslated regions of the cystic fibrosis transmembrane regulator. This screen uncovered a long, evolutionarily conserved motif common to the 3' UTRs of the CFTR and SEC24A transcripts, and shorter, statistically significant motifs unique to either 5' or 3' UTRs. In addition, of the 140 transcripts identified in the screen that encode proteins with known protein interactions, 130 are linked to CFTR through protein interactions. The screen identified genes that are known to be involved in lung fibrosis, the inflammatory response of cystic fibrosis and sensitivity to Pseudomonas aeruginosa infections. The bioinformatic analysis of untranslated regions should prove to be a powerful adjunct to other tools for predicting pathways and relevant interactions.