GLmodel: Fonction to add interaction variable and to fit a group lasso...

In vstanislas/GGEE: R Package for the Group Lasso Gene-Gene Eigen Epistasis (G-GEE) Method

Description

The GLmodel fonction first add interaction variable to a genetic data set structured by group and then fit a group lasso model with an adaptive ridge cleaning approach.

Usage

GLmodel(X, Y, XIntTrain, XIntTest, idSubs, interLength, listGenesSNP,
  nlambda = 100, limitLambda = 30, lambda.cri = "min")

Arguments

X	a genotype matrix where columns are genetic markers and rows are samples.
Y	phenotype values can be logical (1/0) or numericGLmodel.
interLength	a vector that indicate the number of interaction variable for each gene couple.
listGenesSNP	list containing the names of genetic marker for each group.
nlambda	length of the grid of lambda values (100 by default).
limitLambda	number of lambda values among the grid to consider for the cross validation (30 by default).
lambda.cri	criteria for lambda selection are "min" or "oneSE" ("min" by default).
XBet	a matrix where columns are interaction variables and rows are samples.

Value

Returns a list including:

res_GL.min	a matrix nx1 where row are single marker or interaction variables and column the coefficient values obtain with the lambda criteria chosen.
pval.adj	a matrix where row are single marker or interaction variables for which res_GL.min is no equal to zero and column the pvalues obtained after the cleaning procedure and adjusted with Benjamini & Hochberg correction.
vc	a list with result of the cross validation.

vstanislas/GGEE documentation built on May 28, 2021, 12:50 p.m.