Gu X, Li W H
Human Genetics Center, University of Texas at Houston 20334, USA.
Mol Biol Evol. 1996 Dec;13(10):1375-83. doi: 10.1093/oxfordjournals.molbev.a025584.
The statistical properties of the paralinear and LogDet distances under nonstationary nucleotide frequencies were studied. First, we developed formulas for correcting the estimation biases of the paralinear and LogDet distances, i.e., the bias-corrected distance is estimated by dc = d - 2var(d), where d and var(d) are the estimated distance and sampling variance, respectively. The performances of these formulas and the formulas for sampling variances were examined by computer simulation. Second, we developed a method for estimating the variance-covariance matrix of paralinear distances, so that statistical tests of DNA phylogenies can be conducted in the nonstationary case. Third, a new LogDet-based method for testing the molecular clock hypothesis was developed under nonstationary nucleotide frequencies.
研究了非平稳核苷酸频率下 paralinear 距离和 LogDet 距离的统计特性。首先，我们推导了用于校正 paralinear 距离和 LogDet 距离估计偏差的公式，即通过 dc = d - 2var(d) 来估计偏差校正距离，其中 d 和 var(d) 分别是估计距离和抽样方差。通过计算机模拟检验了这些公式以及抽样方差公式的性能。其次，我们开发了一种估计 paralinear 距离方差协方差矩阵的方法，以便在非平稳情况下对 DNA 系统发育进行统计检验。第三，在非平稳核苷酸频率下开发了一种基于 LogDet 的新方法来检验分子钟假设。
