MSSE Test A p-value combination test that combines three independent p-values (element wise if three vectors are given) while prioritizing the first x, as proposed in our pre-print www.biorxiv.org/content/10.1101/2024.09.30.615852. See Details

Let f be a Fisher combination function that maps a set of -log10 p-values to a single -log10 p-value using the classic Fisher combination approach. In a setting where we wish to combine three -log10 p-values – x, y, and z – using f(x,y,z) treats all three -log10 p-values exchangeably. However, in some settings, we may be using y and z to boost the signal in x. In other words, we may be able to safely assume that f(x,y,z) has no chance of being significant is x is not at least somewhat significant. We run into this setting with LOCATER where x corresponds to the single marker test (SMT) -log10 p-value while y and z correspond to tests performed via Stable Distillation and QForm respectively. In this context, MSSE gains power over Fisher combination by ignoring cases where y and/or z would be significant without x being somewhat significant.

Usage

msse.test(x, y, z, test.1.solo = TRUE)

Arguments

x

vector of -log10 p-values

y

vector of -log10 p-values

z

vector of -log10 p-values

test.1.solo

logical, if FALSE ignore situations where x might be significant by itself and focus statistical power on cases where x is only significant in combination with y and/or z. By default = TRUE.

Value

a vector of -log10 p-values