中文摘要：

以细菌和古菌基因组5’UTR序列作为研究对象，分析在5’UTR的3个不同阅读框架中三联体AUG的分布，发现无论是细菌还是古菌基因组都在阅读框1中有非常明显的AUG缺失（depletion）。AUG的缺失表明在起始密码子上游的AUG很可能会对基因的翻译起始产生影响。分析得知：绝大部分的AUG都是以uORF（upstream open readingframe）的形式出现的，uAUG（upstreamAUG）的数量很少，特别是在阅读框1中，而且在细菌基因组的阅读框1中uAUG较多地出现在了含有SD序列的基因上游。比较发现，uAUG引导的序列在同义密码子使用上的偏好性较真正的编码序列差，这可能表明细菌和古菌在同义密码子使用上的偏好性也是决定基因准确地翻译起始的重要因素之一。

英文摘要：

Set three reading frames artificially in 5＇ untranslated region （5＇ UTR） of bacterial and archaeal genomes and analyse the frequencies distribution of triplet AUG in the three reading frames respectively. There is an obviously AUG depletion in reading frame 1 in both bacterial and archaeal genomes. AUG depletion may be an indication that upstream AUG can probably disturb the translation initiation of bacteria and archaea. The authors found that majority of AUGs in the upstream sequences have in-frame stop codons in the upstream sequences and form upstream open reading frame （uORF）, only a fraction of AUGs in the upstream have no in-frame stop codon in the upstream sequences, especially in reading frame 1. The authors call them upstream AUGs （uAUGs）. In bacterial genomes, more uAUGs in reading frame 1 can be found in the genes having a Shine-Dalgarno sequence. Further analyses show synonymous codon usage bias is weaker of the sequences leading by uAUGs than the codon sequences. It may be an indication that the synonymous codon usage bias is an important determinant of translation initiation accurately in bacteria and archaea.