iNOTE (version 1.0)

itegs: Integrative Total Effect of a Gene Set Test

Description

A variance-component based testing procedure to test for the total effect of methylation loci and mRNA expression across a network after specifying an underlying disease risk model which applies to all genes in the gene set of interest.

Usage

itegs(iCPG, iGE, iY, iX, imodel = "mgc", iapprox = "pert", 
  i.omniseed = NA, no.pert = 1000, gsp.emp = TRUE)

Arguments

iCPG

A list of J CpG matrices with dimensions nx\(p_{j}\).

iGE

An nxJ matrix of gene expression values.

iY

An nx1 dichotomous outcome vector.

iX

An nxr covariate matrix.

imodel

The specified disease risk model for the whole gene set: 'mgc', 'mg', or 'm'.

iapprox

The preferred method of gene-set p-value calculation: 'pert' or 'davies'

i.omniseed

A Jx1 seed vector.

no.pert

If using 'pert', the no. perturbations per gene level test; defaults to 1000.

gsp.emp

If using 'pert': The method of calculating the perturbation based p-value: empirical (TRUE) or parametric (FALSE).

Value

A list with the following components:

p

The p-value for the joint, integrative total effect test for the gene set under a pre-specified disease risk model for the whole gene set.

iapprox

The method of gene-set p-value calculation: 'pert' or 'davies'.

imodel

The user-specified disease risk model for the whole gene set.

p.emp

If using 'pert', and if gsp.emp is set to 'FALSE', returns the empirical p-value in addition to the approximated p-value.

gsp.emp

A logical value to indicate the method of calculation for the perturbation based p-value: TRUE for empirical, FALSE for parametric.

References

Chu S.H. and Huang Y-T. (2017) Integrative genomic analysis of biological gene sets with applications in lung cancer. (Revise and Resubmit)

Examples

data(X); data(Y); data(CPG); data(GE)
itegs(iCPG=CPG, iGE=GE, iY=Y, iX=X, imodel='mgc', iapprox='pert', gsp.emp=FALSE);
itegs(iCPG=CPG, iGE=GE, iY=Y, iX=X, imodel='mgc', iapprox='davies');