Go Back

A Structure-Based Statistic for Identifying Coding Regions in mRNAs

Jack E. Tabaska and Michael Q. Zhang

Cold Spring Harbor Laboratory, PO Box 100, Cold Spring Harbor NY 11724

An expanding body of evidence shows that the higher-order structure of the 5' and 3' untranslated regions (UTRs) of an mRNA can have profound effects on the molecule's translation, processing, and stability. We are developing ways to use this information to improve automated detection of terminal exons in genomic sequences. Described herein is a structural potential statistic, or SPS. The SPS measures the possibility that a base is involved in a stable base-pairing interaction with any other nearby base. SPS scores are insensitive to heterogeneity in mRNA secondary structure, allowing us to study the bulk structural features of these molecules. Our results show a sharp and significant decrease in average SPS scores from the 5' UTR to the coding region of an mRNA, and, conversely, a marked increase in scores from the coding region to the 3' UTR. The SPS scores of the mRNA coding region also exhibit a 3-base periodicity which is detectable by spectral methods. These data suggest that our structure potential statistic may be useful for detecting protein-coding regions within genomic sequences.

Go Back