This R package toolkit is at the center of my PhD dissertation research. It fits a hierarchical model to RNA-sequencing data and other high-dimensional count data in fully Bayesian fashion via Markov chain Monte Carlo (MCMC). We accelerated the MCMC using massively parallel, the machinery for which is encapsulated in separate helper packages. The fbseqCUDA package implements GPU-accelerated MCMC with tens of thousands of simultaneous threads, and the fbseqOpenMP implements the same parallelism on a smaller scale using OpenMP. The publication of the method is pending, and my coauthors are Drs. Jarad Niemi and Dan Nettleton of the Iowa State University Department of Statistics.


This R package is for anyone who uses R to analyze multiple datasets in multiple ways. For a common set of use cases, workflowHelper makes it much easier to set up and maintain a reproducible workflow in the midst of heavy development. For the full story and motivation, check out the blog post.


This R package implements parallel computing for Rich Fitzjohn's remake package. For the full story and motivation, check out the blog post.


When I'm setting up a reproducible workflow in R, I don't run the full scaled-up workflow right away. I first run a downsized version to test and debug. This package helps with the downsizing. For more, see the blog post.