Methods for matrix factorization based on Empirical Bayes Matrix Factorization. The name of the package, "flashr," comes from "Factors and Loadings by Adaptive SHrinkage in R".
Please consider using the newer R package, flashier, which contains most of the features of flash, plus many important improvements, including a new, user-friendly interface, more modeling options, faster algorithms, and more detailed statistical outputs.
Copyright (c) 2017-2018, Matthew Stephens and Wei Wang.
All source code and software in this repository are made available under the terms of the BSD 3-Clause License. See the LICENSE file for the full text of the license.
If you find that this R package is useful for your work, please cite our paper:
W. Wang and M. Stephens, 2018. Empirical Bayes matrix factorization. arXiv:1802.06931.
Follow these steps to quickly get started using flashr
.
R
install.packages("devtools")
library(devtools)
install_github("stephenslab/flashr",build_vignettes = TRUE)
This command should automatically retrieve and install the ashr
and ebnm
packages from GitHub (and possibly other packages). If
it does not, install ashr
and ebnm
separately using devtools:
R
install_github("stephens999/ashr")
install_github("stephenslab/ebnm")
Note: If you are interested in attempting to reproduce the results in the Wang and Stephens (2018) manuscript, the flashr release that most closely matches the package used in the paper is version 0.4-10. This release can be installed by running the following in R:
R
install_github("stephenslab/flashr@v0.4-10")
Optionally, install MOSEK and the Rmosek package, for faster model fitting. See the ashr GitHub repository for details.
Run a few toy examples illustrating the flash
function:
R
example("flash")
R
vignette("flash_intro")
R
vignette("flash_advanced")
flashr
package.pkgdown::build_site(mathjax = FALSE)
in R to build the website
using pkgdown. Make sure you are
connected to the Internet while running these commands.This software was developed by Matthew Stephens, Wei Wang, Jason Willwerscheid and Peter Carbonetto at the University of Chicago.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.