linnykos/eSVD2: eSVD2 for performing the eSVD-DE for cohort-wide differential expression analysis

This package is for the eSVD-DE, a way to test for differential expression for single-cell RNA-seq data when many donors/individuals are sequenced, and you wish to test for differential expression at the individual-level (i.e., among individuals), not at cell-level. Specifically, this method uses the eSVD matrix factorization to pool information among genes together, and then carefully adjusts the denoised gene expression to not inflate the Type-1 error. The eSVD matrix factorization (which is the backbone of the eSVD-DE method) assumes that each row and column of a matrix is associated with a low-dimensional vector, and eSVD estimates an embedding for each cell with respect to a hierarchical model where the inner product between the row's and column's latent vectors is the natural parameter of a one-parameter exponential family random variable.

install.packages("remotes")
remotes::install_github("linnykos/eSVD2")