ashr - Methods for Adaptive Shrinkage, using Empirical Bayes

The R package 'ashr' implements an Empirical Bayes approach for large-scale hypothesis testing and false discovery rate (FDR) estimation based on the methods proposed in M. Stephens, 2016, "False discovery rates: a new deal", <DOI:10.1093/biostatistics/kxw041>. These methods can be applied whenever two sets of summary statistics---estimated effects and standard errors---are available, just as 'qvalue' can be applied to previously computed p-values. Two main interfaces are provided: ash(), which is more user-friendly; and ash.workhorse(), which has more options and is geared toward advanced users. The ash() and ash.workhorse() also provides a flexible modeling interface that can accommodate a variety of likelihoods (e.g., normal, Poisson) and mixture priors (e.g., uniform, normal).

Last updated 6 months ago

12.11 score 80 stars 15 packages 728 scripts 4.1k downloads

mashr - Multivariate Adaptive Shrinkage

Implements the multivariate adaptive shrinkage (mash) method of Urbut et al (2019) <DOI:10.1038/s41588-018-0268-8> for estimating and testing large numbers of effects in many conditions (or many outcomes). Mash takes an empirical Bayes approach to testing and effect estimation; it estimates patterns of similarity among conditions, then exploits these patterns to improve accuracy of the effect estimates. The core linear algebra is implemented in C++ for fast model fitting and posterior computation.

Last updated 15 days ago

11.02 score 88 stars 3 packages 616 scripts 482 downloads

susieR - Sum of Single Effects Linear Regression

Implements methods for variable selection in linear regression based on the "Sum of Single Effects" (SuSiE) model, as described in Wang et al (2020) <DOI:10.1101/501114> and Zou et al (2021) <DOI:10.1101/2021.11.03.467167>. These methods provide simple summaries, called "Credible Sets", for accurately quantifying uncertainty in which variables should be selected. The methods are motivated by genetic fine-mapping applications, and are particularly well-suited to settings where variables are highly correlated and detectable effects are sparse. The fitting algorithm, a Bayesian analogue of stepwise selection methods called "Iterative Bayesian Stepwise Selection" (IBSS), is simple and fast, allowing the SuSiE model be fit to large data sets (thousands of samples and hundreds of thousands of variables).

Last updated 4 months ago

9.97 score 179 stars 6 packages 568 scripts 1.6k downloads

mixsqp - Sequential Quadratic Programming for Fast Maximum-Likelihood Estimation of Mixture Proportions

Provides an optimization method based on sequential quadratic programming (SQP) for maximum likelihood estimation of the mixture proportions in a finite mixture model where the component densities are known. The algorithm is expected to obtain solutions that are at least as accurate as the state-of-the-art MOSEK interior-point solver (called by function "KWDual" in the 'REBayes' package), and they are expected to arrive at solutions more quickly when the number of samples is large and the number of mixture components is not too large. This implements the "mix-SQP" algorithm, with some improvements, described in Y. Kim, P. Carbonetto, M. Stephens & M. Anitescu (2020) <DOI:10.1080/10618600.2019.1689985>.

Last updated 11 months ago

8.67 score 11 stars 23 packages 78 scripts 5.3k downloads

EbayesThresh - Empirical Bayes Thresholding and Related Methods

Empirical Bayes thresholding using the methods developed by I. M. Johnstone and B. W. Silverman. The basic problem is to estimate a mean vector given a vector of observations of the mean vector plus white noise, taking advantage of possible sparsity in the mean vector. Within a Bayesian formulation, the elements of the mean vector are modelled as having, independently, a distribution that is a mixture of an atom of probability at zero and a suitable heavy-tailed distribution. The mixing parameter can be estimated by a marginal maximum likelihood approach. This leads to an adaptive thresholding approach on the original data. Extensions of the basic method, in particular to wavelet thresholding, are also implemented within the package.

Last updated 7 years ago

8.58 score 4 stars 14 packages 53 scripts 42k downloads

ebnm - Solve the Empirical Bayes Normal Means Problem

Provides simple, fast, and stable functions to fit the normal means model using empirical Bayes. For available models and details, see function ebnm(). A detailed introduction to the package is provided by Willwerscheid and Stephens (2023) <arXiv:2110.00152>.

Last updated 5 months ago

8.29 score 12 stars 1 packages 110 scripts 351 downloads

fastTopics - Fast Algorithms for Fitting Topic Models and Non-Negative Matrix Factorizations to Count Data

Implements fast, scalable optimization algorithms for fitting topic models ("grade of membership" models) and non-negative matrix factorizations to count data. The methods exploit the special relationship between the multinomial topic model (also, "probabilistic latent semantic indexing") and Poisson non-negative matrix factorization. The package provides tools to compare, annotate and visualize model fits, including functions to efficiently create "structure plots" and identify key features in topics. The 'fastTopics' package is a successor to the 'CountClust' package. For more information, see <doi:10.48550/arXiv.2105.13440> and <doi:10.1186/s13059-023-03067-9>. Please also see the GitHub repository for additional vignettes not included in the package on CRAN.

Last updated 17 days ago

8.10 score 74 stars 1 packages 624 scripts 291 downloads

varbvs - Large-Scale Bayesian Variable Selection Using Variational Methods

Fast algorithms for fitting Bayesian variable selection models and computing Bayes factors, in which the outcome (or response variable) is modeled using a linear regression or a logistic regression. The algorithms are based on the variational approximations described in "Scalable variational inference for Bayesian variable selection in regression, and its accuracy in genetic association studies" (P. Carbonetto & M. Stephens, 2012, <DOI:10.1214/12-BA703>). This software has been applied to large data sets with over a million variables and thousands of samples.

Last updated 1 years ago

4.83 score 2 packages 142 scripts 287 downloads