citcdf

Overview

citcdf is a package to perform conditional independence testing using empirical conditional cumulative distribution function estimations.

The main function of the package is cit(). It uses an asymptotic test (for large sample size , or a permutation test for small sample size with the argument method) to perform conditional independence testing.

The approach implemented in this package is detailed in the following article:

Gauthier M, Agniel D, Thiébaut R & Hejblum BP (2020). Distribution-free complex hypothesis testing for single-cell RNA-seq differential expression analysis, BioRxiv doi:10.1101/2021.05.21.445165

Installation

citcdf is available from GitHub:

#install.packages("devtools")
remotes::install_github("sistm/citcdf")

Example

Here is a basic example which shows how to use citcdf with simple generated data.

## Data Generation
X <- data.frame("X1" = as.factor(rbinom(n=100, size = 1, prob = 0.5)))
Y <- data.frame("Y1" = t(replicate(10, ((X$X1==1)*rnorm(n = 50,0,1)) + ((X$X1==0)*rnorm(n = 50,0.5,1)))))

# Hypothesis testing
res_asymp <- cit(exprmat=Y, variable2test=X, test="asymptotic") # asymptotic test
res_perm <- cit(exprmat=Y, variable2test=X, test="permutation",
                         adaptive=TRUE) # adaptive permutation test

– Marine Gauthier, Denis Agniel, Kalidou Ba, Rodolphe Thiébaut & Boris Hejblum

hex illustration by Jérôme Dubois.