生物多样性

• •    

主流分子钟定年方法的原理、误差来源和优化建议

陈旸康1    1  李家亮1  王文韬1  冯端宇2  毛康珊1*   

  1. 1 (四川大学生命科学学院, 生物资源与生态环境教育部重点实验室, 成都 610065)

    2 (四川大学数学学院, 成都 610065)

  • 收稿日期:2020-07-08 修回日期:2020-08-15 出版日期:2020-10-10 发布日期:2020-10-10
  • 通讯作者: 毛康珊

Principles, error sources and optimization suggestions of prevailing molecular dating methods

Yangkang Chen1, Yi Wang1, Jialiang Li1, Wentao Wang1, Duanyu Feng2, Kangshan Mao1*   

  1. 1 Key Laboratory for Bio-resources and Eco-environment (Ministry of Education), College of Life Sciences, Sichuan University, Chengdu, 610065

    2 College of Mathematics, Sichuan University, Chengdu, 610065

  • Received:2020-07-08 Revised:2020-08-15 Online:2020-10-10 Published:2020-10-10
  • Contact: Kangshan Mao

摘要: 近年来,分子钟定年方法得以广泛运用,为宏观生物学研究、尤其是生物多样性及其格局形成历史的相关研究提供了不可或缺且十分详尽的进化时间框架。贝叶斯方法和马尔可夫链蒙特卡罗方法可容纳多维度、多类型的数据和参数设置,因此以BEAST、PAML-MCMCTree等软件为代表的贝叶斯节点标记法逐渐成为分子钟定年方法中的最为广泛使用的类型。贝叶斯框架的优势之一在于其可以利用复杂模型考虑各种不确定性因素,但是该类方法中各类模型和参数的设置都可能引入误差,从而影响进化分化时间估算的可靠性。本文首先介绍了贝叶斯分子钟定年方法的原理和主要类型;随后,以贝叶斯节点标记法为例,重点讨论了分子钟模型、化石标记的选择与放置、化石标记点年龄先验分布等可能引入误差的环节;最后,我们提出了一些利用贝叶斯节点标记分子钟定年方法重建时间树时需要注意的问题,并针对常见的引起节点年龄潜在高估和低估风险的情况作了分析并给出合理化建议,以期提高贝叶斯节点标记分子钟的可靠性。

关键词: 分子钟, 分子定年, 贝叶斯节点标记法, 系统发育, 进化时间框架, 分化时间。

Abstract: In recent years, molecular dating methods have been widely applied, which provides an indispensable and detailed evolutionary timescale for macrobiological research, especially for studies on the evolutionary history of biodiversity and its distribution pattern. Bayesian methods and Markov chain Monte Carlo methods can accommodate multi-dimensional and multi-type data and parameter settings, which have helped the node-dating method implemented in BEAST, PAML-MCMCTree and other software to become the most widely used type of molecular clock methods. One of the advantages of Bayesian framework is that it can use complex models to consider all kinds of uncertainty factors, so as to make relatively accurate estimation of evolutionary divergence time. This synthesis introduces the principle and main types of Bayesian molecular clock methods, and takes Bayesian node-dating method as an example to discuss the potential errors in molecular clock models, selection and placement of fossil calibrating points, and the prior distribution of node calibrations based on fossils, etc. Finally, important issues that should be paid attention to when using Bayesian node-dating of molecular dating methods to reconstruct evolutionary timescale was summarized, and reasonable suggestions for common occasions that can lead to potential overestimation and underestimation bias of the node ages were provided, aiming at improvements of the reliability of Bayesian node-dating methods.

Key words: molecular clock, molecular dating, Bayesian node-dating, phylogeny, evolutionary timescale, divergence age.