The relation between codon usage, base correlation and gene expression level in Escherichia coli and yeast.

Based on the investigation of the relation between gene expression and the usage of synonymous codons, a method of classifying and predicting the gene expression level is proposed which is called the Self-consistent Information Clustering (SCIC). Using the modified Codon Adaption Index (CAI) values, we have accomplished the linear regression analysis on the relation between base composition, base correlation and gene expression level in Escherichia coli and yeast. The assumption of Expression-Enhancing-Network Site (EENS) is proposed, the existence of which can be demonstrated by the linear equations between gene expression and base correlations in a codon, in adjacent codons and in non-adjacent codons. The modes of base correlation of E. coli and yeast which are important to gene expression have been found and listed in this paper.