Single-Cell RNA-Seq Gene Expression Recovery
An implementation of a regularized regression prediction and
empirical Bayes method to recover the true gene expression profile in
noisy and sparse single-cell RNA-seq data. See Huang M, et al (2018)
for more details.
SAVER (Single-cell Analysis Via Expression Recovery) implements a regularized regression prediction and empirical Bayes method to recover the true gene expression profile in noisy and sparse single-cell RNA-seq data.
Installation
You can install the most recent updates of SAVER from github with:
# install.packages("devtools")
devtools::install_github("mohuangx/SAVER")
To install the stable release of SAVER:
# install.packages("devtools")
Reference
For more information about the method, check out the website , tutorial , and paper .
News
SAVER 1.1.1 (September 26, 2018)
Fix bug when inputted mu values are far from observed
Add get.mu
function to get prior mean predictions.
SAVER 1.1.0 (September 2, 2018)
Improve finish time estimation
Allow input of prior predictions
Can set ncores
to set up parallel backend
Update documentation
Make SAVER tutorial vignette
Allow input of sparse matrices
Allow return of estimates matrix only
SAVER 1.0.0 (April 25, 2018)
SAVER 0.4.0 (February 28, 2018)
Implemented faster version of SAVER which is 20-30x faster
No longer necessary to specify parallel=TRUE
Implemented progress bar
SAVER 0.3.1 (January 29, 2018)
Incorporate iterators
package
Implement mean expression cutoff for prediction matrix
Fix minor bugs
SAVER 0.3.0 (September 19, 2017)
Combine parallel output using unlist
instead of Reduce
.
Reduce
is computationally intensive. Might cause issue #1.
Add nlambda option to saver
Incorporate bigmemory
package
Export cor.cells
SAVER 0.2.2 (August 13, 2017)
Rewrote combine function of parallel implementation of saver
Add null model option to saver
SAVER 0.2.1 (August 1, 2017)
Rewrote parallel implementation of SAVER
Added verbose, pred.cells, and predict.time options to saver
SAVER 0.2.0 (July 23, 2017)
Addressed issue #1
Add option to filter genes out for prediction in saver
Reduced saver
output
Created correlation adjustment functions cor.genes
and cor.cells
SAVER 0.1.3 (July 18, 2017)
Addressed issue #1
Remove requirement of saver
object
Created sample.saver
function
More accurate calculation of finish time
Update version number
SAVER 0.1.2 (May 29, 2017)
Added sample.saver
to sample datasets from the posterior distribution.
Documented dataset and package.
Included size factor, number of genes and cells, and names to saver
output.
Created saver
class.
SAVER 0.1.1 (May 9, 2017)
Prerelease version with documentation.