scBayesDeconv: A Julia package for noise deconvolution

This package implements a Bayesian deconvolution method for extracting the distribution of a target signal $T$ from a measured signal $C$ subject to noise $\xi$:

$$C = T + \xi$$

In other words, we have sample sets from $C$ and $\xi$; and we would like to know the distribution of the signal $T$.

What is implemented in the package?

Bayesian Gaussian Mixture Models:

• Finite Gaussian Mixture Models
• Infinite Gaussian Mixture Models (Dirichlet Processes)

Deconvolution Bayesian Gaussian Mixture Models:

• Finite Deconvolution Gaussian Mixture Models
• Infinite Deconvolution Gaussian Mixture Models (Dirichlet Processes)

Installation

The package can be installed in Julia as follows:

using Pkg
Pkg.add(url="https://github.com/dsb-lab/scBayesDeconv.jl#VERSION")

or alternatively, using the Pkg REPL:

pkg> add https://github.com/dsb-lab/scBayesDeconv.jl#VERSION

for the version of interest (if no version is given, the development version will be installed).

Getting started

To explore the basic usage of these models, see the section Artificial convolutions in the documentation.