IntBIO COLLABORATIVE RESEARCH: Integrating fossils, genomics, and machine learning to reveal drivers of Cretaceous innovations in flowering plants

生命之树的特点是短时间的快速创新，在这些创新中，群体以急剧变化的形式出现，并迅速多样化。这种情况在地质历史上发生过多次，包括鸟类和哺乳动物的兴起，植物从水到陆地的过渡，以及开花植物的起源。特别是开花植物的迅速出现和多样化代表了地球生命史上最引人注目的事件之一。然而，尽管这是理解现代生态系统的生态学和进化的基础，但这一事件仍未得到解释，并导致了生物科学中的一个重大挑战——确定什么过程可能导致生命之树的形式和功能发生如此迅速的变化。处理这一问题的一个主要障碍是数据的可得性和分析这些数据的方法。这项研究的目标是使用和开发新的机器学习方法来收集化石和活植物物种的数据，并利用这些数据来帮助开发新技术。这些技术将有助于确定是什么导致了植物的快速变化，从而导致它们在今天的环境中占据主导地位。该项目将对本科生、研究生和博士后进行机器学习方法、进化生物学以及化石和活标本处理技术方面的培训。该项目还将包括为中学生、高中生、本科生和更广泛的研究界提供资源开发和培训。本项目旨在评估创新出现背后的进化过程，以开花植物为例进行研究。具体而言，该项目将研究白垩纪开花植物的辐射，其特征是快速进化，丰富的化石记录，以及具有重大生态意义的创新和谱系的起源。拟议研究的中心目标是(a)使用新颖的机器学习方法生成大型开花植物形态数据集，(b)开发新的建模进化统计方法，以及(c)利用数据收集和方法方面的这些进展来确定导致生命之树中偶发性和快速出现新颖性的过程。总的来说，机器学习技术、形态学数据收集和用于解决进化过程的分析技术的这些新发展将潜在地改变包括生物科学、计算生物学和机器学习在内的几个领域。该项目的大范围和规模，以及其高度整合的性质，创造了解决进化生物学中最重要的长期问题之一的潜力。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The Tree of Life is marked by short periods of rapid innovation where groups emerge with dramatically altered forms and diversify quickly. This has happened multiple times throughout geologic history including with the rise of birds, mammals, the transition of plants from water to land, and the origination of flowering plants. The rapid emergence and diversification of flowering plants in particular represents one of the most remarkable episodes in the history of life on earth. However, while fundamental to understanding the ecology and evolution of modern ecosystems, this episode remains unexplained and leads to one of the grand challenges in the biological sciences – determining what processes may be responsible for such rapid changes in form and function across the Tree of Life. A major impediment to addressing this question is the availability of data and methods for analyzing those data. The goal of this study is to use and develop new machine learning approaches to gathering data for both fossil and living plant species and to use these data to help develop new techniques. These techniques will help identify what contributed to the rapid change in plants that resulted in their dominance in the environment today. This project will train undergraduates, graduate students, and postdoctoral fellows in machine learning methods, evolutionary biology, and techniques for working with both fossil and living specimens. The project will also include resource development and training for middle schoolers, high schoolers, undergraduates, and the broader research community.This project aims to evaluate the evolutionary processes underlying the emergence of innovation, using flowering plants as a case study. Specifically, the project will examine Cretaceous radiations of flowering plants characterized by rapid evolution, a rich fossil record, and the origin of innovations and lineages of great ecological significance. The central goals of the proposed research are to (a) generate a large morphological dataset for flowering plants using novel machine learning methods, (b) develop new statistical methods for modeling evolution, and (c) use these advances in data collection and methods to identify the processes that led to the episodic and rapid emergence of novelty across the Tree of Life. Collectively, these new developments in machine learning techniques, morphological data collection, and analytical techniques for addressing evolutionary processes will be potentially transformative to several fields including the biological sciences, computational biology, and machine learning. The large scope and scale of this project, together with its highly integrative nature, creates the potential to address one of the most important standing questions in evolutionary biology.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.