Elevated substitution rates estimated from ancient DNA sequences

Ancient DNA sequences are able to offer valuable insights into molecular evolutionary processes, which are not directly accessible via modern DNA. They are particularly suitable for the estimation of substitution rates because their ages provide calibrating information in phylogenetic analyses, circumventing the difficult task of choosing independent calibration points. The substitution rates obtained from such datasets have typically been high, falling between the rates estimated from pedigrees and species phylogenies. Many of these estimates have been made using a Bayesian phylogenetic method that explicitly accommodates heterochronous data. Stimulated by recent criticism of this method, we present a comprehensive simulation study that validates its performance. For datasets of moderate si ze, i t produces accurate estimates of rates, while appearing robust to assumpti ons about demographic history. We then analyse a large collection of 749 ancient and 727 modern DNA sequences from 19 species of animals, plants and bacteria. Our new estimates confirm that the substitution rates estimated from ancient DNA sequences are elevated above long-term phylogenetic levels.

古DNA序列（Ancient DNA sequences）能够为分子演化过程提供极具价值的研究洞见，而这类过程无法通过现代DNA直接观测得到。古DNA序列尤其适用于替换速率（substitution rates）的估算，因其携带的年代信息可为系统发育分析（phylogenetic analysis）提供校准依据，从而规避了选取独立校准位点（calibration points）这一棘手任务。基于此类数据集得到的替换速率通常偏高，介于通过家系（pedigrees）与物种种系系统发育（species phylogenies）估算得到的速率区间内。此类速率估算多采用一种可显性适配异时数据（heterochronous data）的贝叶斯系统发育方法（Bayesian phylogenetic method）。鉴于近期针对该方法的批评意见，本研究开展了一项全面的模拟实验以验证其性能表现。对于中等规模的数据集，该方法可得到精准的速率估算结果，且对种群历史（demographic history）的相关假设具备较强稳健性。随后，本研究对涵盖19种动物、植物与细菌的749条古DNA序列及727条现代DNA序列的大型数据集进行了分析。我们得到的全新估算结果证实，古DNA序列的替换速率高于长期系统发育层面的速率水平。