Collaborative Research: ABI Development: Improving the stability, usability, and speed of the RevBayes platform for phylogenetic analysis

所有物种都通过一棵未知的系谱树相互联系：“生命之树”。生命树的一小部分的谱系树被生物学家称为‘系统发生学’；今天对系统发生学的兴趣比过去任何时候都要大。在流行病学和分子生物学等领域，系统发育是至关重要的，在这些领域中，它们被用来跟踪传染病的传播，在分子生物学中，它们被用来了解分子途径是如何发展的。通过比较不同生物体的特性，可以发现“生命之树”。传统上，系统发育是通过比较不同物种的解剖特征来重建的。最近，来自单个基因，甚至整个基因组的DNA序列信息被比较，以估计系统发育。该项目将继续开发生物学家用来重建系统发育的RevBayes计算机程序。该程序将来自感兴趣的生物体的比较信息作为输入。该程序的输出是解释比较数据的最佳系统发育的概率。具体来说，该项目将提高程序的速度、可用性和可靠性。RevBayes是生物学家广泛使用的Mr Bayes程序的继任者，用于估计系统发育。然而，RevBayes项目与贝耶斯的做法大相径庭。RevBayes程序实现了一种类似R的语言来描述统计模型。模型在计算机存储器中表示为图，其中图的顶点是参数，边表示参数之间的依赖关系。RevBayes使用马尔可夫链蒙特卡罗来逼近参数的后验概率分布。本项目将在几个方面对RevBayes程序进行重大改进：(1)将实施单元和集成测试，以提高程序的可靠性；(2)将更好地利用计算资源，如多核或GPU，以提高程序的速度；(3)将开发跨平台的图形用户界面，以提高程序的可用性；(4)输出将被捆绑，以提高程序的重现性；以及(5)程序将与其他软件程序一起工作，以提高其互操作性。最后，与会者将主持几个研讨会，其中将描述系统发育理论，重点是使用RevBayes将该理论应用于现实世界的问题。源代码可以在http://revbayes.github.io/about.htmlThis的Github网站上找到，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智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

All species are related to one another through an unknown genealogical tree: the `Tree of Life.' Genealogical trees for small parts of the Tree of Life are called `phylogenies' by biologists; there is more interest in phylogenies today than at any time in the past. Phylogenies are crucial in fields such as epidemiology where they are used to track the spread of infectious disease and molecular biology where they are used to understand how molecular pathways developed. The `Tree of Life' can be discovered by comparing the characteristics of different organisms. Traditionally, phylogenies were reconstructed by comparing anatomical traits across different species. More recently, DNA sequence information from individual genes, or even full genomes, are compared to estimate phylogenies. This project will continue the development of the RevBayes computer program that biologists use to reconstruct phylogenies. The program takes as input the comparative information from the organisms of interest. The output of the program is the probabilities of the best phylogenies that explain the comparative data. Specifically, the project will improve the speed, usability, and reliability of the program.RevBayes is the successor of the MrBayes program which is widely used by biologists to estimate phylogeny. However, the RevBayes program represents a significant departure from MrBayes. The RevBayes program implements an R-like language to describe statistical models. The model is represented in computer memory as a graph in which the vertices of the graph are the parameters and the edges represent the dependencies between parameters. RevBayes uses Markov chain Monte Carlo to approximate the posterior probability distribution of parameters. This project will improve the RevBayes program in several signficant ways: (1) unit and integration testing will be implemented to improve the reliability of the program; (2) computational resources, such as multiple cores or GPUs, will be taken better advantage of, to improve the speed of the program; (3) a cross-platform graphical user interface will be developed to improve the usability of the program; (4) output will be bundled to improve the reproducibility of the program; and (5) the program will be made to work with other software programs, to improve its interoperability. Finally, several workshops will be hosted by the participants in which phylogenetic theory will be described with emphasis of application of the theory to real-world problems using RevBayes. The source code can be found on the Github site at http://revbayes.github.io/about.htmlThis 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.