eagle.vem: Run the variational EM algorithm of EAGLE.

In davidaknowles/eagle: Environment-ASE through Generalized LinEar modeling

Description

Run the variational EM algorithm of EAGLE. The specific design matrices determine whether this is for the null or alternative models.

Usage

eagle.vem(alt,n,x,settings)

Arguments

alt	list (over exonic SNPs) of alternative read counts
n	list (over exonics SNPs) of total read counts
x	list of design matrices for each exonic SNP
settings	list of settings. Create using eagle.settings() and then customize.

Value

coeffs	List of regression coefficients at each exonic SNP.
mlPerIteration	(Approximate) marginal likelihood calculated at every EM iteration
log.likelihoods	Per locus marginal likelihoods (used for the likelihood ratio test)
rep.shape	Shape of the variational posterior on the random effect precision at each locus
rep.shape	Rate of the variational posterior on the random effect precision at each locus
rep.global.shape	Shape of the prior on the random effect precisions ($a$ parameter in the manuscript)
rep.global.rate	Rate of the prior on the random effect precisions ($a$ parameter in the manuscript)

Author(s)

David A. Knowles

davidaknowles/eagle documentation built on May 14, 2019, 9:35 p.m.