AI 中文总结
该研究综述并推广全功能生物的进化信息论模型，开发选择的统计检验，扩展形成可变相互作用理论TVI，为量化生物实体及其相互作用的进化变化提供统一框架。
AI 中文摘要
我们综述、阐明并推广了近期提出的全功能生物（holobiont）的进化信息论模型，该模型利用杰弗里斯散度（Jeffreys divergence）量化进化变化，并将其划分为宿主、微生物组分以及宿主-微生物组关联的贡献。基于这些划分，我们开发统计检验以识别观测到的信息变化是否归因于对宿主类型、微生物组分状态或特定宿主-微生物组合的选择作用。随后，我们将该框架扩展至受组内和组间选择的多组分群体，最终形成我们称为可变相互作用理论（Theory of Variable Interactions, TVI）的通用层级公式。在该公式中，生物单元可包含相互作用的组分，且自身可形成更高层级的集合，从而允许将信息变化递归划分为任意数量组织层级上的边际和关联组分。该框架涵盖了此前研究的公地悲剧模型、集合与多组分全功能生物选择模型，以及作为进一步特例的非随机交配和性选择的信息模型。因此，TVI为量化生物实体及其相互作用的进化变化提供了统一的信息论框架，其在各层级保留了局部统计维度，而更高层级则引入了额外的关系维度。
英文摘要
We review, clarify, and generalize a recently proposed evolutionary information-theoretic model of the holobiont, in which evolutionary change is quantified using Jeffreys divergence and partitioned into contributions from the host, microbial components, and host-microbiome associations. Building on these partitions, we develop statistical tests to identify whether observed informational change is attributable to selection acting on host types, microbial-component states, or particular host-microbiome combinations. We then extend the framework to multicomponent groups subject to within- and between-group selection and, ultimately, to a general hierarchical formulation that we call the Theory of Variable Interactions (TVI). In this formulation, biological units may contain interacting components and may themselves form higher-level sets, allowing informational change to be partitioned recursively into marginal and association components across an arbitrary number of organizational levels. The framework encompasses previously studied models of the tragedy of the commons and of aggregate and multicomponent holobiont selection, and, as a further specialization, informational models of non-random mating and sexual selection. TVI therefore provides a unified information-theoretic framework for quantifying evolutionary change in both biological entities and their interactions, with local statistical dimensionality preserved across hierarchical levels while higher levels introduce additional relational dimensions.
Comments74 pages, 2 figures, 4 tables
