Skip to content
/ lute Public

R package for deconvolution simulation and optimization.

Notifications You must be signed in to change notification settings

metamaden/lute

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

R build status

lute

Authors: Sean Maden, Stephanie Hicks

lute is a framework for deconvolution experiments.

Installation

From GitHub

Install lute from GitHub by running the following in an R session:

devtools::install("metamaden/lute")

Framework overview

The lute R package supports bulk transcriptomics deconvolution experiments. It gets its name from the word "deconvolute," the problem of predicting cell amounts from cell mixtures. Users can use the lute() function to select for markers and perform deconvolution either with or without a cell size scale factor normalization (see ?lute for details). For normalizations, data may be incorporated from the cellScaleFactor package and passed to the s argument for transformation of cell reference expression data prior to deconvolution.

Deconvolution methods supported

lute supports standard access to deconvolution algorithms with the deconvolution() generic function and affiliated class definitions. The below algorithms are currently supported or have bee tested using these specialized classes. They are listed with links to their main repos as well as conda YML scripts to install them with dependencies.

The following algorithms are included in the current build of lute:

  • NNLS : The nnlsParam class supports the nnls::nnls implementation of the NNLS deconvolution algorithm (url; yml).

  • Bisque : The bisqueParam class supports the BisqueRNA::ReferenceBasedDecomposition implementation of the Bisque deconvolution algorithm (url; yml).

The following algorithms have been supported and tested using lute's specialized classes, but they have not been included in the latest build.

  • MuSiC : The musicParam class supports the MuSiC::music.basic() implementation of the MuSiC deconvolution algorithm (url; yml).

  • MuSiC2 : The music2Param class supports both the MuSiC::music2_prop() and MuSiC2::music2_prop() implementations of the MuSiC2 deconvolution algorithm (url; yml).

  • EPIC : The epicParam class supports the EPIC::EPIC() implementation of the EPIC deconvolution algorithm (url; yml).

  • DeconRNASeq : The deconrnaseqParam class supports the DeconRNASeq::DeconRNASeq() implementation of the DeconRNASeq deconvolution algorithm (url; yml).

  • SCDC : The scdcParam class supports the SCDC::SCDC_prop() implementation of the SCDC deconvolution algorithm (url; yml).

Conda environments

To run specific deconvolution methods and algorithms, lute contains conda YML scripts. For example, set up an environement to run the NNLS algorithm with the following:

conda env create -f ./lute/inst/yml/nnls.yml

Acknowledgements

We acknowledge the following individuals for their helpful feedback and suggestions on improving this project: Louise Huuki-Myers

About

R package for deconvolution simulation and optimization.

Resources

Stars

Watchers

Forks

Packages

No packages published