Codon stabilization coefficient as an index for prediction of codon bias and its potential applications.

Abstract

Different methods of mRNA half-life measurements are available, but genome-wide measurements of mRNA half-life in yeast showed a weak correlation between the methods. Moreover, when we compared mRNA half-life determined by these methods with other cellular measurements such as mRNA and protein abundance low correlation was found. To clarify this matter, we analyzed mRNA half-life datasets from nine different groups to determine the most accurate method of measurement. Since codon optimality is one of the significant determinants of mRNA stability, we used the codon stabilization coefficient (CSC) as a reference for mRNA half-life measurement accuracy. After CSC calculation for each dataset, we find strong positive correlations between the CSC from some datasets with other parameters that reflect codon optimality such as tRNA abundance and ribosome residence time. By the use of CSC parameter, we observed that most genes contain non-optimal codons and that codon bias exists toward optimized and non-optimized genes. We also observed that stretches of non-optimal are not randomly distributed since it causes impacts on translation.