GSFA

GSFA (Guided Sparse Factor Analysis) is an R package (accelerated by Rcpp) that performs sparse factor analysis and differential gene expression discovery simultaneously on single-cell RNA-seq with multiplexed CRISPR screen (e.g. CROP-seq, Perturb-seq) data.

The method assumes that the perturbation of a target gene affects certain latent factors (gene modules), which in turn alters the expression of individual genes. By fitting a Bayesian hierarchical model using Gibbs sampling, GSFA infers the latent factors and their associations with the perturbations in a joint statistical framework. It also summarizes the effects of a perturbation on individual genes as the sum of effects mediated by all the factors.

Provided with a normalized gene expression matrix and a perturbation matrix (of which cells contain which type of gRNAs) a single-cell CRISPR screening experiment, GSFA can
(1) identify genetically controlled factors that are associated with the perturbation;
(2) interpret the biological meanings of factors (gene modules) through gene ontology/pathway enrichment analysis thanks to the sparse gene weights on factors;
(3) detect differentially expressed genes under each genetic perturbation by thresholding the local false sign rate (LFSR).

Citing this work

If you find the GSFA package or any of the source code in this repository useful for your work, please cite:

Yifan Zhou, Kaixuan Luo, Lifan Liang, Mengjie Chen and Xin He. A new Bayesian factor analysis method improves detection of genes and biological processes affected by perturbations in single-cell CRISPR screening. Nature Methods. (2023). doi: 10.1038/s41592-023-02017-4. PMID: 37770710.

License

All source code and software in this repository are made available under the terms of the MIT license.

Installation

Github

To install the development version of the GSFA package from Github, run this in R:

install.packages("devtools")
devtools::install_github("xinhe-lab/GSFA", build_vignettes = TRUE)

Note that installing the package will require a C++ compiler setup that is appropriate for the version of R installed on your computer.

Source

If you have cloned the repository locally, you can install the package with the install_local function from devtools. Assuming you are in the local GSFA repository, run this code in R to install the package:

devtools::install_local(build_vignettes = TRUE)

Note that installing the package will require a C++ compiler setup that is appropriate for the version of R installed on your computer.

Docker

If you are familiar with Docker, this R package and all its C++ and R dependencies built on debian stable have been containerized in a docker image. Run the code below to pull the docker image of the release version of GSFA. For more details on the dependencies, please see Dockerfile in this repository.

docker pull gradonion/gsfa:latest

Once pulled, run the docker image (by modifying the docker run command below) and launch RStudio Server locally at localhost:8787 to run the vignettes or your own analyses.

docker run \
    -d \
    -e DISABLE_AUTH=true \
    -v /your_local_directory/:/home/rstudio/projects/ \
    -p 8787:8787 \
    gradonion/gsfa:latest

Using the package

Please see this package vignette for using GSFA on a simulated example:

library(GSFA)
vignette("gsfa_intro")

For guidance on using GSFA to analyze and interpret real single-cell CRISPR screen data, please refer to the code and examples in this repository.

Credits

The GSFA package is developed by Yifan Zhou from the He Lab and Chen Lab at the University of Chicago.