Pole-aware Cauchy (ACAT) transform of p-values

tan((0.5 - p) * pi), computed so that a p-value at or near either endpoint gives a large finite number rather than an infinity. Near the poles the tangent loses precision, so the limits \(1/(p\pi)\) and \(-1/\{(1-p)\pi\}\) are substituted, which is what Liu's own ACAT implementation does.

Usage

acat_term(p = NULL, log_p = NULL, log_1mp = NULL)

Arguments

p

Numeric vector of p-values in \([0, 1]\).

log_p

Numeric vector of \(\log(p_k)\), an alternative to p.

log_1mp

Numeric vector of \(\log(1 - p_k)\).

Value

Numeric vector the same length as the input.