Phylo Lab Software
AutoParts: Bayesian inference of phylogeny from partitioned sequence data
Developers: Brian Moore, John Huelsenbeck, Sebastian Höhna
It is well established that the evolutionary process may vary across the sites of nucleotide alignments; ‘process heterogeneity’ is an increasingly pervasive feature of datasets composed of gene regions sampled from multiple loci and/or different genomes. Inference of phylogeny from these data demands that we adequately model the underlying process heterogeneity; failure to do so can lead to biased estimates of phylogeny and other parameters. Traditionally, process heterogeneity has been accommodated under a ‘mixed-model’ approach, where sites are first assigned to data subsets based on relevant prior information, and then the phylogeny and other model parameters are estimated under the resulting mixed model. AutoParts provides an alternative approach for accommodating process heterogeneity that treats the partition scheme as a random variable using a Dirichlet process prior model, where the phylogeny is estimated by integrating over all possible process partitions for the specified data subsets. AutoParts provides a practical solution for inferring phylogeny from complex multi-gene/omic datasets, allowing discovery of novel process partitions that may more effectively balance error variance and estimation bias, while rendering phylogenetic inference more robust to process heterogeneity by virtue of integrating estimates over all possible partition schemes. AutoParts is a C++ program that has been carefully documented (with a comprehensive user manual, step-by-step tutorials), and helper programs, such as the AutoPlots R package (to automate graphical summaries of the results).
RevBayes: A flexible framework for Bayesian inference of phylogeny
Developers: Sebastian Höhna, Michael J. Landis, Tracy A. Heath, Bastien Boussau, Nicolas Lartillot, Brian R. Moore, John P. Huelsenbeck, Fredrik Ronquist
Programs for Bayesian inference of phylogeny currently implement a unique and fixed suite of models. Consequently, users of these software packages are simultaneously forced to use a number of programs for a given study, while also lacking the freedom to explore models that have not been deemed interesting by the developers of those programs. RevBayes seeks to address these problems. RevBayes is the successor of MrBayes; however, these programs do not share a single line of code. RevBayes is entirely based on probabilistic graphical models, a powerful generic framework for specifying and analyzing statistical models. Phylogenetic graphical models can be specified interactively in RevBayes, piece by piece, using a new succinct and intuitive language called Rev. The primary strength of RevBayes is the simplicity to design, specify and implement new and complex (comparative) phylogenetic models. The graphical-model framework also provides pedagogical advantages: it removes the typical black box so that users can visualize the model they are specifying. This transparency will improve the understanding of phylogenetic models in our field, and will motivate the search for improvements to existing methods by brazenly exposing the model choices that we make to critical scrutiny. RevBayes has been exceptionally well documented with 17 detailed tutorials, and a comprehensive (250-page) user manual.
TESS: Bayesian inference of lineage diversification rates
Developers: Sebastian Höhna, Michael R. May, Brian R. Moore
Many fundamental questions in evolutionary biology entail estimating rates of lineage diversification (speciation – extinction) that are modeled using birth-death branching processes. We leverage recent advances in branching-process theory to develop a flexible Bayesian framework for specifying diversification models—where rates are constant, vary continuously, or change episodically through time—and implement numerical methods to estimate parameters of these models from molecular phylogenies, even when species sampling is incomplete. We enable both statistical inference and efficient simulation under these models. We also provide robust methods for comparing the relative and absolute fit of competing branching-process models to a given tree, thereby providing rigorous tests of biological hypotheses regarding patterns and processes of lineage diversification. This versatile software package has been exceptionally well documented (with detailed tutorials in a comprehensive [103-page] user manual)
BONSAI: Automating the analysis of Bayesian MCMC output
Developers: Michael R. May, Sebastian Höhna, Brian R. Moore
Bayesian inference of phylogeny relies on numerical methods—Markov chain Monte Carlo (MCMC) algorithms—to approximate the joint posterior probability density of the model parameters. Studies typically involve a large number of MCMC simulations; i.e., replicate simulations are required for each of the candidate phylogenetic models under consideration. Accordingly, manual processing of MCMC output quickly becomes tedious. Moreover, relevant MCMC diagnostics are distributed across a number of separate software tools. To address this issue, we developed BONSAI (Bayesian Output Needs Semi-Automated Inspection). BONSAI automates a comprehensive suite of MCMC diagnostics and generates a detailed report that flags possible issues for subsequent inspection by the user. These reports can also serve as supplemental materials for publications to allow reviewers and editors to easily assess the reliability of MCMC simulations. We are currently preparing the BONSAI software package for release (expected release date: January 2016).