Skip to contents
devil 0.99.0
- Initial submission to Bioconductor.
- Implements negative binomial GLM differential expression with cluster-robust sandwich covariance estimators.
- Accepts raw count matrices, SingleCellExperiment, and SummarizedExperiment objects as input via
fit_devil().
- Supports CPU parallelism via BiocParallel and optional GPU acceleration (CUDA).
- Multiple overdispersion estimation strategies: MOM, iterative MLE, and Poisson (no overdispersion).
- Multiple size factor normalization methods: normed_sum, psinorm, and edgeR TMM.
