Fig. 1

Factorial single-cell latent variable model. a f-scLVM is based on a variant of factor analysis, decomposing the matrix of single-cell gene expression profiles into factors and weights. Gene sets from pathway databases are used to annotate a subset of factors, with the remainder enabling the activation of unannotated factors. Unannotated factors include both sparse, likely biological, factors and dense factors, which are more likely to explain confounding sources of variation. b The fitted model can be used for different downstream analyses, including i) identification of biological drivers; ii) visualization of cell states; iii) data-driven adjustment of gene sets; and iv) estimation of residual expression dataset, thereby adjusting for unwanted variation and confounding effects