CoMM

CoMM to dissecting genetic contributions to complex traits by leveraging regulatory information.

lmm_pxem(y, w, x, maxIter)

Arguments

y	gene expression vector.
w	covariates file for eQTL data.
x	normalized genotype (cis-SNPs) matrix for eQTL.
maxIter	maximum iteration (default is 1000).

Value

List of model parameters

Details

lmm_pxem fits the linear mixed model. (n < p)

Author

Jin Liu, jin.liu@duke-nus.edu.sg

Examples

L = 1; M = 100; rho =0.5
n1 = 350; n2 = 5000;
X <- matrix(rnorm((n1+n2)*M),nrow=n1+n2,ncol=M);

beta_prop = 0.2;
b = numeric(M);
m = M * beta_prop;
b[sample(M,m)] = rnorm(m);
h2y = 0.05;
b0 = 6;

y0 <- X%*%b + b0;
y  <- y0 + (as.vector(var(y0)*(1-h2y)/h2y))^0.5*rnorm(n1+n2);
h2 = 0.001;
y1 <- y[1:n1]
X1 <- X[1:n1,]
y = y1;

mean.x1 = apply(X1,2,mean);
x1m = sweep(X1,2,mean.x1);
std.x1 = apply(x1m,2,sd)
x1p = sweep(x1m,2,std.x1,"/");
x1p = x1p/sqrt(dim(x1p)[2])
w1 = matrix(rep(1,n1),ncol=1);

fm0 = lmm_pxem(y, w1,x1p, 100)