mleBeta: Maximum likelihoood estimate of beta.

In ipw012/RIVER: R package for RIVER (RNA-Informed Variant Effect on Regulation)

Description

mleBeta computes maximum likelihoood estimate of beta (parameters between FR (functionality of regulatory variant) and G (genomic annotations); multivariate logistic regression).

Usage

mleBeta(Feat, FuncRv, costs)

Arguments

Feat	Genomic features (G)
FuncRv	Soft-assignments of FR from E-step
costs	Candidate penalty parameter values for L2-regularization within logistic regression.

Value

MLE of beta

Warning

To input a vector of candidate penalty values makes glmnet faster than to input a single penalty value

Author(s)

Yungil Kim, [e-mail]

See Also

glmnet

Examples

dataInput <- getData(filename=system.file("extdata", "simulation_RIVER.gz",
        package = "RIVER"), ZscoreThrd=1.5)
Feat <- scale(t(Biobase::exprs(dataInput))) # genomic features (G)
Out <- as.vector(as.numeric(unlist(dataInput$Outlier))-1) # outlier status (E)
theta.init <- matrix(c(.99, .01, .3, .7), nrow=2)
costs <- c(100, 10, 1, .1, .01, 1e-3, 1e-4)
logisticAllCV <- glmnet::cv.glmnet(Feat, Out, lambda=costs, family="binomial",
        alpha=0, nfolds=10)
probFuncRvFeat <- getFuncRvFeat(Feat, logisticAllCV$glmnet.fit, logisticAllCV$lambda.min)
posteriors <- getFuncRvPosteriors(Out, probFuncRvFeat, theta=theta.init)
logistic.cur <- mleBeta(Feat, FuncRv=posteriors$posterior, costs)

ipw012/RIVER documentation built on March 8, 2020, 7:54 p.m.