Bayesian Modelling for Label Free MS Based Proteomics
StanProt is a R package that provides Bayesian models for Label
Free MS based proteomics data. It allows a.o. to model the data missingness
process simultaneously to the MS intensities.
On the one hand, StanProt proposes some S4 classes to standardise the access,
manipulation and interpretation of Bayesian models generated with the Stan
MCMC Engine (exact inference). See here.
On the other hand, StanProt provides also support for approximate
inference, i.e. MFVB approximation (using built-in code), and Laplace
approximation (using also Stan as en engine).
The StanProt code is provided under GPL license version 3.0 or
higher.
The documentation,
including the manual pages and the vignettes, are distributed under a CC BY-SA
4.0 license.
If you use StanProt in your research, please cite the following paper:
Hauchamps, Philippe. A Bayesian Approach combining Peptide Intensity and Missingness Modelling to analyse Label Free Mass Spectrometry based Proteomics Data. Faculté des sciences, Université catholique de Louvain, 2021. Prom. : Lambert, Philippe ; Gatto, Laurent. You can find a link to the manuscript here.
It is recommended to install first the rstan package and to make sure it works.
See the rstan quick start guide.
As soon as this is done, and the installation is verified by running a test model,
the next step is to install the StanProt package from the github repo:
devtools::install_github("https://github.com/UCLouvain-CBIO/StanProt")
See R scripts in examples directory.