Skip to contents

Calculates the 1-degree-of-freedom weighted burden statistic \(T_\text{burden} = w^T Z / \sqrt{w^T Rw}\) of a given window of SNPs. Under the null hypothesis of no association in the window, \(T\) asymptotically follows a standard normal distribution \(N(0,1)\).

Usage

t_burden(w, Z, R)

Arguments

w

Numeric vector. Discovery weights (e.g., marginal effect sizes from the discovery cohort) for the SNPs in the window.

Z

Numeric vector. Marginal Z-scores from the target cohort for the SNPs in the window. Must be the same length as w.

R

Numeric matrix. Linkage Disequilibrium (LD) correlation matrix for the SNPs in the window. Dimensions must be length(w) x length(w).

Value

A single numeric value representing the burden test statistic. Returns 0 if the variance denominator is non-positive or non-finite.