A joint hierarchical model for the number of cases and deaths due to COVID-19 across the boroughs of Montreal

Montreal is the epicentre of the COVID-19 pandemic in Canada with highest number of deaths. The cumulative numbers of cases and deaths in the 33 areas of Montreal are modelled through bivariate hierarchical Bayesian models using Poisson distributions. The Poisson means are decomposed in the log scale as the sums of fixed effects and latent effects. The areal median age, the educational level, and the number of beds in long-term care homes are included in the fixed effects. To explore the correlation between cases and deaths inside and across areas, three bivariate models are considered for the latent effects, namely an independent one, a conditional autoregressive model, and one that allows for both spatially structured and unstructured sources of variability. As the inclusion of spatial effects change some of the fixed effects, we extend the Spatial+ approach to a Bayesian areal set up to investigate the presence of spatial confounding.

Revticulate: An R framework for interaction with RevBayes

1: Phylogenetic methods are increasingly complex. Researchers need to make many choices about how to model different aspects of the data appropriately. It is increasingly common to deploy hierarchical Bayesian models in which different data types may be described by different processes. This necessitates tools to help users understand model assumptions more clearly.2: We describe the package \code{Revticulate}, which provides an R-based interface to the software RevBayes. RevBayes is a Bayesian phylogenetics program that implements an R-like computing language, but does not interface with R itself. Revticulate was designed to allow communication between an R session, and all of its associated capabilities, such as plotting and simulation, and a RevBayes session.3: Revticulate can be used to copy objects from RevBayes into R. We provide several usage examples demonstrating how objects, such as such as random variables drawn from probability distributions and phylogenetic trees, can be generated in RevBayes. We then show how these objects can be used with R's phylogenetic ecosystem to plot a phylogenetic tree, or with base R functions to simulate the behavior of a particular probability. 4: Revticulate is a broadly useful software. Revticulate can be used alongside popular document preparation packages, such as Knitr and pkgdown to generate attractive reports, tutorials, and websites. This means that researchers who are looking to communicate their work in RevBayes can do that very easily using Revticulate, enabling rapid generation of reproducible research outputs.

Principled Decision-Making Workflow with Hierarchical Bayesian Models of High-Throughput Dose-Response Measurements

We present a case study applying hierarchical Bayesian estimation on high-throughput protein melting-point data measured across the tree of life. We show that the model is able to impute reasonable melting temperatures even in the face of unreasonably noisy data. Additionally, we demonstrate how to use the variance in melting-temperature posterior-distribution estimates to enable principled decision-making in common high-throughput measurement tasks, and contrast the decision-making workflow against simple maximum-likelihood curve-fitting. We conclude with a discussion of the relative merits of each workflow.

Principled decision-making workflow with hierarchical Bayesian models of high throughput dose-response measurements

We present a case study applying hierarchical Bayesian estimation on high throughput protein melting point data measured across the tree of life. We show that the model is able to impute reasonable melting temperatures even in the face of unreasonably noisy data. Additionally, we demonstrate how to use the variance in melting temperature posterior distribution estimates to enable principled decision-making in common high throughput measurement tasks, and contrast the decision-making workflow against simple maximum-likelihood curve fitting. We conclude with a discussion of the relative merits of each workflow.

A Risk Assessment for Ozone Regulation Based on Statistical Rollback

In environmental studies, it is important to assess how regulatory standards for air pollutants affect public health. High ozone levels contribute to harmful air pollutants. The EPA regulates ozone levels by setting ozone standards to protect public health. It is thus crucial to assess how various regulatory ozone standards affect non-accidental mortality related to respiratory deaths during the ozone season. The original rollback approach provides an adjusted ozone process under a new regulation scenario in a deterministic fashion. Herein, we consider a statistical rollback approach to allow for uncertainty in the rollback procedure by adopting the quantile matching method so that it provides flexible rollback sets. Hierarchical Bayesian models are used to predict the potential effects of different ozone standards on human health. We apply the method to epidemiologic data.