DM: Compute the distance-to-median statistic

In MarioniLab/scran: Methods for Single-Cell RNA-Seq Data Analysis

Compute the distance-to-median statistic

Description

Compute the distance-to-median statistic for the CV2 residuals of all genes

Usage

DM(mean, cv2, win.size=51)

Arguments

mean	A numeric vector of average counts for each gene.
cv2	A numeric vector of squared coefficients of variation for each gene.
win.size	An integer scalar specifying the window size for median-based smoothing. This should be odd or will be incremented by 1.

Details

This function will compute the distance-to-median (DM) statistic described by Kolodziejczyk et al. (2015). Briefly, a median-based trend is fitted to the log-transformed cv2 against the log-transformed mean using runmed. The DM is defined as the residual from the trend for each gene. This statistic is a measure of the relative variability of each gene, after accounting for the empirical mean-variance relationship. Highly variable genes can then be identified as those with high DM values.

Value

A numeric vector of DM statistics for all genes.

Author(s)

Jong Kyoung Kim, with modifications by Aaron Lun

References

Kolodziejczyk AA, Kim JK, Tsang JCH et al. (2015). Single cell RNA-sequencing of pluripotent states unlocks modular transcriptional variation. Cell Stem Cell 17(4), 471–85.

Examples

# Mocking up some data
ngenes <- 1000
ncells <- 100
gene.means <- 2^runif(ngenes, 0, 10)
dispersions <- 1/gene.means + 0.2
counts <- matrix(rnbinom(ngenes*ncells, mu=gene.means, size=1/dispersions), nrow=ngenes)

# Computing the DM.
means <- rowMeans(counts)
cv2 <- apply(counts, 1, var)/means^2
dm.stat <- DM(means, cv2)
head(dm.stat)

MarioniLab/scran documentation built on March 7, 2024, 1:45 p.m.