Abstract
The details in the mechanism of the translation initiation in eukaryotes are still unclear. Though Kozak proposed that the certain sequence, GCC(A/G)CCaugG, controls the efficiency of the translation initiation, she could use the restricted number of genes because genomes had not been decoded. Moreover, it has been reported that the specific sequences are different among each eukaryote species. Previously we have invented a method to represent the statistical biases in nucleotide appearance at each position around the initiation codon. In this analysis, we applied this method to 48 eukaryote species and conducted comparative analyses among them. Then, a strong bias of A/G at the position of -3 and that of C at +5 (the second position in the second codon) were commonly observed on all the species examined. By using a cluster analysis among the patterns of the biases in nucleotide appearance, we revealed that evolutionarily closely related species show similar patterns of the nucleotide biases. We also detected an intriguing three-base period in the 5' untranslated regions for several species. The sequences identified in this study are thought to control the translation initiation.