corr.calibration: Calibration for the correlation test

In gnonadd: Various Non-Additive Models for Genetic Associations

Calibration for the correlation test

Description

This function finds appropriate scaling_parameters for the kappa_calc function by means of bootstrapping.

Usage

corr.calibration(qt1, qt2, reps = 10000)

Arguments

qt1	A numeric vector.
qt2	A numeric vector.
reps	The number of repeates we want to perform for each sample size

Value

A list with the values:

* bias_scale, a value to determine the bias for the correlation estimators * weight_scale, a value to determine how much weight we assign to each correlation estimator * safe_weight_scale, same as weight scale, but larger. Using this weight decreases the chance of type 1 error, with the cost of statistical power.

Examples

n_val <- 10000
Q <- MASS::mvrnorm(n = n_val, mu = c(0,0), Sigma = matrix(c(1,0.3,0.3,1),
                   nrow = 2, ncol = 2))
qt1_val <- Q[,1]
qt2_val <- Q[,2]
res <- corr.calibration(qt1_val, qt2_val, 10)

gnonadd documentation built on Sept. 22, 2023, 5:07 p.m.