devil

devil is an R package for differential expression analysis in single-cell RNA sequencing (scRNA-seq) data. It supports both single- and multi-patient experimental designs, implementing robust statistical methods to identify differentially expressed genes while accounting for technical and biological variability.

Key features are:

1. Flexible experimental design support (single/multiple patients)
2. Robust statistical testing framework
3. Efficient implementation for large-scale datasets

Installation

You can install the current version of devil from GitHub with:

devtools::install_github("caravagnalab/devil")

Example

This is a basic example which shows you how to fit the expression for a single gene observed in 1000 cells.

library(devil)
y <- t(as.matrix(rnbinom(1000, 1, .1)))
fit <- devil::fit_devil(input_matrix = y, design_matrix = matrix(1, ncol = 1, nrow = 1000), verbose = T, size_factors = NULL, overdispersion = "MOM")
#> Initialize theta
#> Initialize beta
#> Fitting beta coefficients
#> Fit overdispersion (mode = MOM)
test <- devil::test_de(fit, contrast = c(1))

---

Copyright and contacts

Giulio Caravagna, Giovanni Santacatterina. Cancer Data Science (CDS) Laboratory.