Selvar: Estimation of heritability in high dimensional sparse linear...

In EstHer: Estimation of Heritability in High Dimensional Sparse Linear Mixed Models using Variable Selection

Description

This function selects active components in sparse linear mixed models in order to estimate heritability. The selection allows us to reduce the size of the data sets which improves the accuracy of the estimations. Our package also provides a confidence interval for the estimated heritability.

Usage

Selvar(Y,Z,X,thresh_vect,nb_boot=80,nb_repli=50,CI_level=0.95,nb_cores=1)

Arguments

Y	Vector of observations of size n.
Z	Matrix with genetic information of size n x N.
X	Matrix of fixed effects of size n x d.
thresh_vect	Vector of thresholds in the stability selection step: the higher the threshold, the smallest the set of selected components.
nb_boot	Number of subsamples of Y to apply our bootstrap technique. The value by default is 80.
nb_repli	Number of replications in the stability selection. The value by default is 50.
CI_level	Level of the confidence interval for the estimation of the heritability. The value by default is 0.95.
nb_cores	Number of cores of the computer. It is used for parallelizing the computations. The value by default is 1.

Value

heritability	Estimation of the heritability
CI_up	Upper bound of the confidence interval for the estimated heritability
CI_low	Lower bound of the confidence interval for the estimated heritability
selec_ind	Indexes of the columns of the selected components

Author(s)

Anna Bonnet and Celine Levy-Leduc

Examples

library(EstHer)
data(Y)
data(W)
data(X)
Z=scale(W,center=TRUE,scale=TRUE)
res=Selvar(Y,Z,X,thresh_vect=c(0.7,0.8,0.9),nb_boot=80,nb_repli=50,CI_level=0.95,nb_cores=1) 
res$heritability
res$CI_low
res$CI_up

EstHer documentation built on May 2, 2019, 8:49 a.m.