Suzuki Yoshiyuki
Institute of Molecular Evolutionary Genetics and Department of Biology, The Pennsylvania State University, USA.
Genes Genet Syst. 2007 Aug;82(4):341-60. doi: 10.1266/ggs.82.341.
Natural selection operating on amino acid substitution at single amino acid sites can be detected by comparing the rates of synonymous (r(S)) and nonsynonymous (r(N)) nucleotide substitution at single codon sites. Amino acid substitutions can be classified as conservative or radical according to whether they retain the properties of the substituted amino acid. Here methods for comparing the rates of conservative (r(C)) and radical (r(R)) nonsynonymous substitution with r(S) at single codon sites were developed to detect natural selection operating on these substitutions at single amino acid sites. A method for comparing r(C) and r(R) at single codon sites was also developed to detect biases toward these substitutions at single amino acid sites. Charge was used as the property of the amino acids. In a computer simulation, false-positive rates of these methods were always < 5%, unless termination sites were included in the computation of the numbers of sites and estimates of transition/transversion rate ratio were highly biased. The frequency of detection of natural selection operating on conservative substitution was almost independent of the presence of natural selection operating on radical substitution, and vice versa. Natural selection operating specifically on conservative and radical substitution was detected more efficiently by comparing r(S) with r(C) and r(S) with r(R) than by comparing r(S) with r(N). These methods also appeared to be robust against the occurrence of recombination during evolution. In an analysis of class I human leukocyte antigen, negative selection operating on conservative substitution, but not positive selection operating on radical substitution, was observed at some of the codon sites with r(R) > r(C), suggesting that r(R) > r(C) may not necessarily be an indicator of positive selection operating on radical substitution.
通过比较单个密码子位点上同义核苷酸替换率(r(S))和非同义核苷酸替换率(r(N)),可以检测作用于单个氨基酸位点氨基酸替换的自然选择。氨基酸替换可根据是否保留被替换氨基酸的特性分为保守型或激进型。本文开发了在单个密码子位点上比较保守型(r(C))和激进型(r(R))非同义替换率与r(S)的方法,以检测作用于单个氨基酸位点这些替换的自然选择。还开发了一种在单个密码子位点比较r(C)和r(R)的方法,以检测单个氨基酸位点上对这些替换的偏好。电荷被用作氨基酸的特性。在计算机模拟中,除非在计算位点数量时包含终止位点且转换/颠换率比值估计存在高度偏差,否则这些方法的假阳性率始终<5%。作用于保守替换的自然选择的检测频率几乎与作用于激进替换的自然选择的存在无关,反之亦然。与比较r(S)与r(N)相比,通过比较r(S)与r(C)以及r(S)与r(R)能更有效地检测专门作用于保守和激进替换的自然选择。这些方法似乎也对进化过程中重组的发生具有稳健性。在对人类白细胞抗原I类的分析中,在一些r(R)>r(C)的密码子位点观察到作用于保守替换的负选择,但未观察到作用于激进替换的正选择,这表明r(R)>r(C)不一定是作用于激进替换的正选择的指标。