Class: SkatOptimalTest

stats~SkatOptimalTest()

Optimal sequence kernel association test (SKAT).

The following papers detail the method:

Original SKAT optimal test paper, utilizing genotypes instead of covariance matrices: https://doi.org/10.1016/j.ajhg.2012.06.007 Meta-analysis of SKAT optimal test, and use of covariance matrices: https://doi.org/10.1016/j.ajhg.2013.05.010

Constructor

new SkatOptimalTest()

Source:

Extends

  • AggregationTest

Methods

(static) weights(mafs, a, b)

Calculate typical SKAT weights using beta density function.

Parameters:
Name Type Default Description
mafs Array.<number>

Array of minor allele frequencies.
a number 1

alpha defaults to 1.
b number 25

beta defaults to 25.
Source:

run(u, v, w, mafs, rhos) → {Array.<Number>}

Calculate optimal SKAT test.

This code is based partly on rvtests' implementation (https://git.io/fjQEs) which uses genotypes instead of scores/covariances, and also on the MetaSKAT R-package (https://git.io/fjQEZ).

Parameters:
Name Type Description
u Array.<Number>

Vector of score statistics (length m, number of variants).
v Array.<Number>

Covariance matrix of score statistics (m x m).
w Array.<Number>

Weight vector (length m, number of variants). If weights are not provided, they will be calculated using the default weights() method of this object.
mafs Array.<Number>

A vector of minor allele frequencies. These will be used to calculate weights if they were not provided.
rhos Array.<Number>

A vector of rho values, representing the weighting between burden and SKAT statistics.
Source:
Returns:

SKAT p-value.

Type
Array.<Number>