Macroevolutionary Rates of Species Interactions: Approximate Bayesian Inference from Cophylogenies

Understanding the macroevolutionary dynamics of species interactions such as parasitisms, commensalisms, and mutualisms is an important goal in evolutionary ecology. To this end, statistical inference from extant cophylogenetic systems holds immense potential. However, such inference cannot yet handle the speciation-extinction dynamics that occur simultaneously in the host and symbiont clades on the same timescale. Here we present an Approximate Bayesian Computation (ABC) approach that, while taking into account host and symbiont extinction, infers rates of four types of speciation from a cophylogenetic system: (i) host speciation, (ii) symbiont speciation without host-switching, (iii) symbiont speciation with host-switching, and (iv) cospeciation. The new ABC approach relies on a novel design of summary statistics combining both size-based (i.e., tree sizes) and size-free summary statistics (i.e., the normalized distribution of Branch Length Differences - BLenD) of the cophylogeny. Convergence analyses show that the combined design of summary statistics outperforms size-based or size-free summary statistics alone – achieving satisfactory accuracy in detecting rate heterogeneity between the four types of speciation. Our ABC approach allows the user to infer the predominant mode of speciation within a given cophylogenetic system. The approach is demonstrated with an application to a cophylogenetic dataset of commensalism, in which beetles of one genus mimic those of another. In this system, we identify host speciation as the predominant process with the fastest rate (in events per unit time) among all four types of speciation (i.e., 4.3-5.1 times faster than the median among all four types of speciation). Understanding how and why different types of speciation may predominate in different cophylogenetic systems can have implications for various areas in ecology and evolution such as host conservatism, trait-driven diversification, pathogen spillover risk, and parasite extinction risk. This new approach highlights the need and potential for future efforts to compile time-calibrated cophylogenies.

解析寄生、偏利共生、互利共生等物种互作的宏观演化动力学，是进化生态学领域的重要研究目标。为此，基于现存共演化系统的统计推断具备巨大应用潜力，但当前此类推断尚无法处理宿主与共生体支系在同一时间尺度下同步发生的物种形成-灭绝动力学过程。

我们提出一种近似贝叶斯计算（Approximate Bayesian Computation, ABC）方法，该方法在纳入宿主与共生体灭绝因素的同时，可从共演化系统中推断四类物种形成过程的速率：（1）宿主物种形成；（2）无宿主转换的共生体物种形成；（3）伴随宿主转换的共生体物种形成；（4）共物种形成。

该新型ABC方法创新性地设计了摘要统计量，将共系统发生树的基于大小的统计量（即树的规模）与无大小依赖的统计量（即枝长差异的标准化分布，缩写为BLenD）相结合。收敛性分析结果表明，联合使用两类统计量的设计方案优于单独使用基于大小或无大小依赖的统计量，在检测四类物种形成过程间的速率异质性时可获得令人满意的准确率。

本ABC方法可帮助研究者推断特定共演化系统中占主导地位的物种形成模式。我们通过一个偏利共生类群的共演化数据集验证了该方法的有效性：该数据集涉及两个甲虫属之间的拟态关系。在此系统中，我们发现宿主物种形成是四类物种形成过程中速率最快（以单位时间内事件数计）的主导过程，其速率比四类过程的中位数快4.3至5.1倍。

解析不同类型的物种形成为何以及如何在不同共演化系统中占据主导地位，可为进化生态学的多个研究方向提供启示，例如宿主保守性、性状驱动的物种分化、病原体溢出风险以及寄生虫灭绝风险等。该新型方法凸显了未来构建经时间校准的共系统发生树的必要性与研究潜力。