A dimensionality reduction method to facilitate gene selection for targeted gene profiling by learning a sparse gene encoding of single cells

scPNMF is a method to facilitate gene selection for targeted gene profiling by learning a sparse gene encoding of single cells. Compared with existing gene selection methods, scPNMF has two advantages. First, its selected informative genes can better distinguish cell types, with a small number, e.g., < 200 genes. Second, it enables the alignment of new targeted gene profiling data with reference data in a low-dimensional space to help the prediction of cell types in the new data.

scPNMF can be installed from Github with the following code in R:

install.packages("devtools")
library(devtools)

install_github("JSB-UCLA/scPNMF")

For detailed info on scPNMF method and applications, please check out the package vignettes, or with the following code in R:

install_github("JSB-UCLA/scPNMF", build_vignettes = TRUE)
browseVignettes("scPNMF")

Any questions or suggestions on scPNMF are welcomed! Please report it on issues, or contact Dongyuan Song (dongyuansong@ucla.edu) or Kexin Li (aileenlikexin@outlook.com).

Dongyuan Song, Kexin Li, Zachary Hemminger, Roy Wollman, Jingyi Jessica Li, scPNMF: sparse gene encoding of single cells to facilitate gene selection for targeted gene profiling, Bioinformatics, Volume 37, Issue Supplement_1, July 2021, Pages i358–i366, https://doi.org/10.1093/bioinformatics/btab273