IIBR RoL: Collaborative Research: A Rules Of Life Engine (RoLE) Model to Uncover Fundamental Processes Governing Biodiversity

由于数据的数量和种类越来越多，现在存在着前所未有的机会来促进对控制生命多样性和动态的生物规则的理解。 当人类系统破坏这些动态时，了解生物多样性动态的这一目标是在关键时刻实现的。然而，从数据中获取理解所需的科学和计算工具仍然缺失。 这些工具需要可供广大用户群体使用，从而促进参与和创新。 该项目将（1）建立一个针对生物多样性——物种丰度、遗传、功能和系统发育等多个方面的计算模型； （2）通过检验三个示范系统中生物多样性产生和维持的主要假设来使用和完善该模型； （3）通过构建开源平台来准备不同的数据源并运行模型，使模型可供科学界使用； (4) 创建教学有效的课程和研讨会，使来自不同背景的学生、研究人员和利益相关者能够了解生物多样性理论以及用数据测试这些理论所需的数据科学工具。生命规则引擎 (RoLE) 模型将是一个机械的、基于模拟的假设检验和数据合成框架，使科学家能够利用多维生物多样性数据生成和检验有关驱动生物多样性模式的过程的假设。 RoLE 模型将应用机器学习新技术，使模型适合高维、跨尺度数据。该模型将模拟基于个体的生态和遗传中性模型的生态进化社区组装构建，并添加非中性、基于性状的竞争和环境过滤。新的物种和性状将通过元群落的长期进化和局部群落的快速进化而出现。当地社区的种群遗传学和物种丰度将通过出生、死亡、移民和突变进行建模。该项目研究团队将通过测试三个生物地理系统中的四个假设的生命规则来完善和说明 RoLE 模型的使用，目前这些系统的多尺度生物多样性数据是可用的。这些假设涉及群落组装中移民与物种形成的相对作用、物种相互作用如何影响多样性、不同的组装历史如何决定物种相互作用的强度以及系统是否/如何达到平衡。项目领导者建立了一个由 14 个合作者组成的网络，其中包括国家生态观测站网络，他们将在其不同的系统中使用 RoLE 模型并推广更广泛的采用。为了进一步减少使用障碍，RoLE模型框架将作为开源软件提供，包括带有标准化元数据输出的R语言Shiny App界面，以促进可重复性和共享。从 RoLE 模型中获得的见解与保护直接相关，例如，群落是否主要通过物种形成或移民聚集，很大程度上决定了它们对人为压力和最佳保护管理的反应。为了鼓励参与定量生物多样性研究，项目负责人将通过圣达菲研究所的复杂性探索者计划开发一门大规模开放的在线课程，使用 RoLE 模型作为交互式教学工具。研究团队还将与 Data Carpentry 和 Software Carpentry 一起提供面对面的数据科学培训研讨会。 RoLE 项目的结果可以在 https://role-model.github.io 上找到。该奖项反映了 NSF 的法定使命，并且通过使用基金会的智力价值和更广泛的影响审查标准进行评估，被认为值得支持。

An unprecedented opportunity to advance understanding of the biological rules that govern the diversity and dynamics of life now exists thanks to the large quantity and variety of data that are becoming increasingly available. This goal of understanding biodiversity dynamics is enabled at a critical moment when human systems are disrupting those very dynamics. However, the scientific and computational tools needed to derive understanding from data are still missing. Such tools need to be accessible to a broad community of users, thereby catalyzing involvement and innovation. This project will (1) build a computational model for multiple aspects of biodiversity-species abundance, genetic, functional, and phylogenetic; (2) use and refine this model by testing major hypotheses about the generation and maintenance of biodiversity in three exemplar systems; (3) make the model accessible to the scientific community by building an open-source platform to prepare diverse data sources and run the model; and (4) create pedagogically effective courses and workshops to enable students, researchers, and stakeholders from many backgrounds to understand biodiversity theory and the data science tools needed to test those theories with data.The Rules of Life Engine (RoLE) model will be a mechanistic, simulation-based hypothesis-testing and data synthesis framework enabling scientists with multi-dimensional biodiversity data to generate and test hypotheses about the processes driving biodiversity patterns. The RoLE model will apply new techniques in machine learning to fit models to high dimensional, cross-scale data. The model will simulate eco-evolutionary community assembly building from individual-based ecological and genetic neutral models with added non-neutral, trait-based competition and environmental filtering. New species and traits will arise through long time scale evolution in the meta-community and rapid evolution in the local community. Population genetics and species abundances in the local community will be modeled through birth, death, immigration, and mutation. The project research team will refine and illustrate the use of the RoLE model by testing four hypothesized rules of life across three bio-geographic systems for which multi-scale biodiversity data are now available. The hypotheses address the relative roles of immigration versus speciation in community assembly, how species interactions influence diversity, how different assembly histories determine the strength of species interactions, and whether/how systems come to equilibrium. The project leaders have established a network of 14 collaborators, including the National Ecological Observatory Network, who will use the RoLE model in their diverse systems and propagate wider adoption. In order to further reduce barriers to use, the RoLE model framework will be made available as open source software, including an R language Shiny App interface with standardized metadata outputs to promote reproducibility and sharing. The insights gained from the RoLE model are of direct relevance to conservation, e.g., whether or not communities are assembled primarily by speciation or immigration strongly determines their response to anthropogenic pressures and optimal conservation management. To encourage participation in quantitative biodiversity research, the project leaders will develop a massively open online course through the Santa Fe Institute?s Complexity Explorer program using the RoLE model as an interactive teaching tool. In conjunction with Data Carpentry and Software Carpentry, the research team will also provide an in-person data science training workshop. Results from the RoLE project can be found at https://role-model.github.io.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.