Conditional Independence Testing with Cumulative Distribution Functions • citcdf

Overview

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

The package has two main entry points: cit_multi() for gene-wise conditional independence testing across many outcomes, and cit_gsa() for gene-set analysis. Both accept test = "asymptotic" (large samples) or test = "permutation" (small samples). The single-outcome workhorses cit_asymp() and cit_perm(), the CCDF estimator ccdf(), and the plotting functions plot_compare_ccdf(), plot.cit_multi() and plot.cit_gsa() are also exported.

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

Gauthier M, Agniel D, Thiébaut R & Hejblum BP (2021). 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 on CRAN:

install.packages("citcdf")

The development version is available from GitHub:

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

Example

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

library(citcdf)

## Data Generation
set.seed(123)
n <- 100
X <- data.frame(X1 = as.factor(rbinom(n = n, size = 1, prob = 0.5)))
Y <- replicate(10, (X$X1 == 1) * rnorm(n) + (X$X1 == 0) * rnorm(n, mean = 0.5))

# Hypothesis testing
res_asymp <- cit_multi(M = data.frame(Y = Y), X = X,
                       test = "asymptotic", parallel = FALSE) # asymptotic test
res_asymp$pvals
#>         raw_pval   adj_pval test_statistic
#> Y.1  0.000126502 0.00126502      30.418136
#> Y.2  0.317402043 0.31740204       4.115804
#> Y.3  0.310454844 0.31740204       4.372046
#> Y.4  0.034961503 0.07955292      11.802747
#> Y.5  0.063473425 0.09067632      10.451354
#> Y.6  0.044804258 0.07955292      12.780768
#> Y.7  0.012650228 0.05834245      15.581473
#> Y.8  0.017502734 0.05834245      14.080686
#> Y.9  0.047731750 0.07955292       8.920593
#> Y.10 0.161833735 0.20229217       6.991416
plot(res_asymp)

plot_compare_ccdf(Y=Y[, 1, drop=FALSE], X=X)

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

hex illustration by Jérôme Dubois.