byExtremality: For array-based DNA methylation, particularly summarized...

In ttriche/sesamizeGEO: the poor man's methylation suite (cf. `sesamizer`)

For array-based DNA methylation, particularly summarized across regions, we can do better (a lot better) than MAD. Since we know that

Description

max(SD(X_j)) if X_j ~ Beta(a, b) < max(SD(X_j)) if X_j ~ Bernoulli(a/(a+b))

Usage

byExtremality(x, k = 500)

Arguments

x	a matrix of beta values, or a GenomicRatioSet
k	how many rows to return (500)

Details

for X having a known mean and SD, hence solvable for a + b by MoM, define

extremality = sd(X_j) / bernoulliSD(mean(X_j))

This function selects the k most extremal rows of x and returns their values.

Value

the most extremal k rows of x (returns the same class as x)

ttriche/sesamizeGEO documentation built on Nov. 12, 2023, 5:42 p.m.