Estimation and sampling for multi-dimensional Hurdle models on a Normal density with applications to single-cell co-expression
A Hurdle model is a modification of a distribution at zero.
This package provides routines to estimate and sample from multivariate Hurdle models on a Normal density.
These are distributions that are conditionally Normal, but with singularities along the coordinate axes, so generalize a univariate zero-inflated distribution.
The main functionality is to estimate graphical models using group-lasso penalized neighborhood selection.
BiocManager::install('amcdavid/HurdleNormal')
or
BiocManager::install('amcdavid/HurdleNormal', dependencies = TRUE)
if you want to compile the vignette or run examples from the paper.
A vignette with an example is located
here
or with vignette('singlecell-networks').
Graphical Models for Zero-Inflated Single Cell Gene Expression