haplo.glm.control {haplo.stats}

Create list of control parameters for haplo.glm

Description

Create a list of control pararameters for haplo.glm. If no parameters are passed to this function, then all default values are used.

Usage

haplo.glm.control(haplo.effect="add", haplo.base=NULL,
                  haplo.min.count=NA, haplo.freq.min=.01,
                  sum.rare.min=0.001, haplo.min.info=0.001, 
                  keep.rare.haplo=TRUE,
                  eps.svd=sqrt(.Machine$double.eps),
                  glm.c=glm.control(maxit=500), 
                  em.c=haplo.em.control())

Arguments

haplo.effect

the "effect" of a haplotypes, which determines the covariate (x) coding of haplotypes. Valid options are "additive" (causing x = 0, 1, or 2, the count of a particular haplotype), "dominant" (causing x = 1 if heterozygous or homozygous carrier of a particular haplotype; x = 0 otherwise), and "recessive" (causing x = 1 if homozygous for a particular haplotype; x = 0 otherwise).

haplo.base

the index for the haplotype to be used as the base-line for regression. By default, haplo.base=NULL, so that the most frequent haplotype is chosen as the base-line.

haplo.min.count

The minimum number of expected counts for a haplotype from the sample to be included in the model. The count is based on estimated haplotype frequencies. Suggested minimum is 5.

haplo.freq.min

the minimum haplotype frequency for a haplotype to be included in the regression model as its own effect. The haplotype frequency is based on the EM algorithm that estimates haplotype frequencies independent of trait.

sum.rare.min

the sum of the "rare" haplotype frequencies must be larger than sum.rare.min in order for the pool of rare haplotypes to be included in the regression model as a separate term. If this condition is not met, then the rare haplotypes are pooled with the base-line haplotype (see keep.rare.haplo below).

haplo.min.info

the minimum haplotype frequency for determining the contribution of a haplotype to the observed information matrix. Haplotypes with less frequency are dropped from the observed information matrix. The haplotype frequency is that from the final EM that iteratively updates haplotype frequencies and regression coefficients.

keep.rare.haplo

TRUE/FALSE to determine if the pool of rare haplotype should be kept as a separate term in the regression model (when keep.rare.haplo=TRUE), or pooled with the base-line haplotype (when keep.rare.haplo=FALSE).

eps.svd

argument to be passed to Ginv for the generalized inverse of the information matrix, helps to determine the number of singular values

glm.c

list of control parameters for the usual glm.control (see glm.control).

em.c

list of control parameters for the EM algorithm to estimate haplotype frequencies, independent of trait (see haplo.em.control).

Value

the list of above components

See Also

haplo.glm, haplo.em.control, glm.control

Examples

# NOT RUN
# using the data set up in the example for haplo.glm,
# the control function is used in haplo.glm as follows
#  > fit <- haplo.glm(y ~ male + geno, family = gaussian,  
#  >          na.action="na.geno.keep",
#  >          data=my.data, locus.label=locus.label,
#  >          control = haplo.glm.control(haplo.min.count=5,
#  >          em.c=haplo.em.control(n.try=1)))