Skip to main content

Table 1 Parameters of the hierarchical gamma-negative binomial (hGNB) model and their interpretations in the context of scRNA-seq data

From: Bayesian gamma-negative binomial modeling of single-cell RNA sequencing data

Parameter Constraint Interpretation
rj rj>0 Expression heterogeneity of genes in sample j
ϕvk \(\sum _{v=1}^{V} \phi _{vk}=1, \phi _{vk}>0\) Gene-latent factor association
θjk θkj>0 Popularity of factor k in sample j
βvp \(\beta _{vp} \in \mathbb {R}\) Impact of cell covariate p on expression of gene v
δjq \(\beta _{vp} \in \mathbb {R}\) Impact of gene covariate q on expression of cell j
  1. The inputs of hGNB are gene counts nvj and vector of cell- and gene-level covariates xj and zv