Local thermodynamic stability scores are well represented by a non-centralstudent's t distribution

Local folding in mRNAs is closely associated with biological functions. rn this study, we reveal the whole distribution of local thermodynamic stabili ty in the complete genome of the poliovirus P3/Leon/37 and the single-stran ded RNA sequences that corresponds to the nucleotide sequence of the comple te genome sequence (1 667 867 bp) of Helicobacter pylori (H. pylori) strain 26695. Local thermodynamic stability in the RNA sequences is measured by t wo standard z-scores, significance score and stability score. To estimate t he distribution of thermodynamic stability, a model based on the non-centra l Student's t distribution has been developed. Significant patterns of extr emes that are either much more stable or unstable than expected by chance a re detected. Our results indicate that the highly stable and statistically more significant folding regions are predominantly in non-coding sequences in the two genome sequences. Moreover, the highly unstable folding regions, on the contrary, are predominantly in the protein coding sequences of H. p ylori.. The observed differences across the complete genomic sequences are statistically very significant by a chi (2)-test. These extreme patterns ma y be useful in searching for target sequences for long-chain antisense RNA and for locating potential RNA functional elements involved in the regulati on of gene expression including translation, mRNA localization and metaboli sm. (C) 2001 Academic Press.