GBJ_objective

Calculates the GBJ objective function at given threshold points. Used both to calculate the
observed GBJ statistic and also to find the boundary points for p-value calculation (through uniroot).
Mostly for internal use.

internal

Offers the Generalized Berk-Jones (GBJ) test for set-based inference in genetic
association studies. The GBJ is designed as an alternative to tests such as Berk-Jones (BJ),
Higher Criticism (HC), Generalized Higher Criticism (GHC), Minimum p-value (minP), and Sequence
Kernel Association Test (SKAT). All of these other methods (except for SKAT) are also implemented
in this package, and we additionally provide an omnibus test (OMNI) which integrates information from each of the tests.
The GBJ has been shown to outperform other tests in genetic association studies when signals
are correlated and moderately sparse. Please see the vignette for a quickstart guide or Sun and Lin
(2017) <arXiv:1710.02469> for more details.

Ryan Sun

Generalized Berk-Jones Test for Set-Based Inference in Genetic
Association Studies

GBJ_objective function

<dl><dt>t_vec</dt>
<dd>A scalar or vector of threshold points (magnitudes of test statistics)
to calculate the objective at.</dd>
<dt>d</dt>
<dd>The total number of test statistics in the set.</dd>
<dt>k_vec</dt>
<dd>If t_vec is not a vector all of the test statistics, let us know which ordered
objective functions we are calculating (can be a vector with the same length as t).</dd>
<dt>pairwise_cors</dt>
<dd>A vector of all d(d-1)/2 pairwise correlations between the test
statistics, where d is total number of test statistics in the set.</dd>
<dt>offset</dt>
<dd>Used if we want to subtract the observed gbj value for uniroot.</dd></dl>

Arguments

GBJ_objective.R — GBJ_objective

<dl>

<dt>t_vec</dt>
<dd>A scalar or vector of threshold points (magnitudes of test statistics)
to calculate the objective at.</dd>


<dt>d</dt>
<dd>The total number of test statistics in the set.</dd>


<dt>k_vec</dt>
<dd>If t_vec is not a vector all of the test statistics, let us know which ordered
objective functions we are calculating (can be a vector with the same length as t).</dd>


<dt>pairwise_cors</dt>
<dd>A vector of all d(d-1)/2 pairwise correlations between the test
statistics, where d is total number of test statistics in the set.</dd>


<dt>offset</dt>
<dd>Used if we want to subtract the observed gbj value for uniroot.</dd>

</dl>

GBJ_objective: GBJ_objective.R

Description

Usage

Value

Arguments

Examples