Conservation of Translation Initiation Sites Based on Dinucleotide Frequency and Codon Usage in Escherichia coli K-12 (W3110): Non-random Distribution of A/T-rich Sequences Immediately

Dinucleotide frequencies are useful for characterizing consensus elements as a minimum unit of nu-cleotide sequence because the neighborhood relations of nucleotide sequences are reflected in dinucleotides. Using a consensus score based on dinucleotide frequencies and intra-species codon usage heterogeneity, denoted by the Z1 parameter, we report the relationship between nucleotide conservation at the translation initiation sites of genes in the Escherichia coli K-12 genome (W3110) and codon usage in its downstream genes. Significant positive correlations were obtained in three regions centered at −13, −4, and +7, which correspond to the Shine-Dalgarno element, the A + T element immediately upstream of the translation initiation site, and the downstream box, respectively. Key words: Escherichia coli K-12 (W3110); translation initiation; codon usage; dinucleotide As genome projects are being completed, genome se-quence data for about 70 species of prokaryotes and eukaryotes are now available. This makes it possible to identify the genomic elements such as codon usage for protein-coding sequences and consensus sequences for transcriptional promoters and terminators as well as translation initiation regions. To understand the ge-nomic elements related to the biological functions of Escherichia coli, we have constructed an E. coli K-12 genome database referred to as GenoBase1,2 which is open to the public through the Interne