Smith Nick G C, Brandström Mikael, Ellegren Hans
Department of Evolutionary Biology, Evolutionary Biology Centre, Uppsala University, Norbyvägen 18D, 752 36 Uppsala, Sweden.
Genomics. 2004 Nov;84(5):806-13. doi: 10.1016/j.ygeno.2004.07.012.
The vast majority of the mammalian genome does not code for proteins, and a fundamental question in genomics is: What proportion of the noncoding mammalian genome is functional? Most attempts to address this issue use sequence comparisons between highly diverged mammals such as human and mouse to identify conservation due to negative selection. But such comparisons will underestimate the true proportion of functional noncoding DNA if there is turnover, if patterns of negative selection change over time. Here we test whether the inferred level of negative selection differs between different pairwise species comparisons. Using a multiple alignment of more than a megabase of contiguous sequence from eight mammalian species, we find a strong negative relationship between inferred levels of negative selection and pairwise divergence using 21 pairwise comparisons. This result suggests that there is a high rate of turnover of functional noncoding elements in the mammalian genome, so measures of functional constraint based on human-mouse comparisons may seriously underestimate the true value.
绝大多数哺乳动物基因组并不编码蛋白质，基因组学中的一个基本问题是：非编码哺乳动物基因组中有功能的部分占比是多少？解决这个问题的大多数尝试都利用人类和小鼠等高度分化的哺乳动物之间的序列比较来识别因负选择而产生的保守性。但是，如果存在更替现象，如果负选择模式随时间变化，那么这种比较将会低估功能性非编码DNA的真实比例。在这里，我们测试了在不同的两两物种比较中，推断出的负选择水平是否存在差异。通过对来自八个哺乳动物物种的超过100万个碱基的连续序列进行多重比对，我们利用21组两两比较发现，推断出的负选择水平与两两差异之间存在强烈的负相关关系。这一结果表明，哺乳动物基因组中功能性非编码元件的更替率很高，因此基于人类与小鼠比较的功能限制测量可能会严重低估真实值。
