Abstract
In eukaryotes, targeting the small ribosomal subunit to the mRNA transcript requires a Kozak sequence at the translation initiation site. Despite the critical importance of the Kozak sequence to regulation of gene expression, there have been no correlation studies between its natural variance and efficiency of translation. Combining bioinformatics analysis with molecular biology techniques, and using zebrafish as a test case, we identify Kozak sequences based on their natural variance and characterize their function in vivo. Our data reveal that while the canonical Kozak sequence is efficient, in zebrafish it is neither the most common nor the most efficient translation initiation sequence. Rather, the most frequent natural variation of the Kozak sequence is almost twice as efficient. We conclude that the canonical Kozak sequence is a poor predictor of translation efficiency in different model organisms. Furthermore, our results provide an experimental approach to testing and optimizing an important tool for molecular biology.