Bayesian Hierarchical Modeling for Label-Free Proteomics.

Baldur

The goal of Baldur is to shine light on your proteomics data. Baldur is a hierarchical Bayesian model that uses an empirical Bayes method to estimate hyperparamters for the variance and measurement specific uncertainty. It then estimates the posterior of the difference in means between different conditions for each peptide/protein/PTM. Finally, it integrates the posterior to estimate the probability of error.

Installation

You can install the development version of baldur from this github or the stable version from CRAN. Importantly, you first need to follow the instructions for installing rstanhttps://github.com/stan-dev/rstan/wiki/RStan-Getting-Started and make sure that is working. Then you can install baldur accordingly:

For the stable release please install from CRAN:

install.packages('baldur')

Or you can install the developmental version of baldur from this github:

devtools::install_github('PhilipBerg/baldur', build_vignettes = T)

Note that Ubuntu operating systems can require pandochttps://pandoc.org/ to compile the vignettes.

For Windows, the developmental version of rstan is sometimes needed to install baldur.

Example

Please see the vignettes for examples vignette('baldur_yeast_tutorial') and vignette('baldur_ups_tutorial').

Reference

Berg, Philip, and George Popescu. “Baldur: Bayesian hierarchical modeling for label-free proteomics exploiting gamma dependent mean-variance trends.” bioRxiv (2023): 2023-05. https://doi.org/10.1101/2023.05.11.540411