itegs: Integrative Total Effect of a Gene Set Test
In iNOTE: Integrative Network Omnibus Total Effect Test

Description Usage Arguments Value Author(s) References Examples

View source: R/itegs.R

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.

1 2	itegs(iCPG, iGE, iY, iX, imodel = "mgc", iapprox = "pert", i.omniseed = NA, no.pert = 1000, gsp.emp = TRUE)

`iCPG`	A list of J CpG matrices with dimensions nxp_{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).

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.

Su H. Chu & Yen-Tsung Huang

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

1
2
3

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');