Alternative translation initiation (ATLI) refers to the existence of multiple translation initiation sites per gene and is a widespread phenomenon in eukaryotes. ATLI is commonly assumed to be advantageous through creating proteome diversity or regulating protein synthesis. We here propose an alternative hypothesis that ATLI arises primarily from nonadaptive initiation errors presumably due to the limited ability of ribosomes to distinguish sequence motifs truly signaling translation initiation from similar sequences. Our hypothesis, but not the adaptive hypothesis, predicts a series of global patterns of ATLI, all of which are confirmed at the genomic scale by quantitative translation initiation sequencing in multiple human and mouse cell lines and tissues. Similarly, although many codons differing from AUG by one nucleotide can serve as start codons, our analysis suggests that using non-AUG start codons is mostly disadvantageous. These and other findings strongly suggest that ATLI predominantly results from molecular error, requiring a major revision of our understanding of the precision and regulation of translation initiation.

中文翻译：

---

哺乳动物替代翻译的启动大多是不自适应的。

替代翻译起始（ATLI）是指每个基因存在多个翻译起始位点，是真核生物中普遍存在的现象。通常认为ATLI通过创建蛋白质组多样性或调节蛋白质合成是有利的。在这里，我们提出了另一种假设，即ATLI主要是由于非适应性启动错误而引起的，大概是由于核糖体区分序列基序的能力有限，这些序列基序真正地指示了翻译起始于相似序列。我们的假设（而非适应性假设）预测了一系列ATLI的整体模式，所有这些模式均已通过在多个人类和小鼠细胞系和组织中进行定量翻译起始测序而在基因组规模上得到了证实。同样，尽管许多不同于一个核苷酸的AUG密码子可以用作起始密码子，但我们的分析表明，使用非AUG起始密码子在大多数情况下是不利的。这些和其他发现强烈表明ATLI主要是由分子错误引起的，需要对我们对翻译起始精度和调控的理解进行重大修订。