one.rGLM.test {rGLM}R Documentation

Return a likelihood ratio like test of haplotype effect based on rGLM approach

Description

This function returns a likelihood ratio like test of haplotype effect comparing two nested models fit with rGLM function.

Usage

one.rGLM.test(nonSNPcolumns,hypoDat, family=binomial, allelic=TRUE, pooling.tol = 0.05, maxit=200, tol=0.001, lambda=0, trace=FALSE) 

Arguments

nonSNPcolumns number of columns that do not contain genotype information in hypoDat data set
hypoDat data set that contains the response variable and genotypes on SNPs
family binomial, poisson, gaussian or freq are supported, default=binomial
allelic TRUE if single-locus SNP genotypes are in allelic format and FALSE if in genotypic format; default is TRUE.
pooling.tol pooling tolerance; by default set to 0.05
maxit maximum number of iterations of the EM algorithm; default=50
tol convergence tolerance in terms of either the maximum difference in parameter estimates between iterations or the maximum relative difference in parameter estimates between iterations, which ever is larger. Default is 0.001.
lambda tuning parameter lambda
trace indicates whether or not some internal results should be printed; default is FALSE.

Value

data data set used in the association test
pre.hapassoc.data data list obtained from pre.hapassoc function
LRTstat an likelihood ratio like statistic
regrH1 a list obtained from rGLM function based on the alternative hypothesis, i.e., a full model is used in the regression
regrH0 a list obtained from rGLM function based on the null hypothesis, i.e., a formula as "affected~1." is used in the regression
H1.converged indicate if regrH1 is converged
H0.converged indicate if regrH0 is converged
na.resason indicate if LRTstat exists and reason that LRTstat is unavailable; If LRTstat exists, the value is "conv", while the values are "na.pre.hapassoc", "unconv.H1", "unconv.H0", "unconv.H1.H0" when LRTstat is unavailable due to prepare the data list by pre.hapassoc function, H1.converged=FALSE, H0.converged=FALSE, and H1.converged=H0.converged=FALSE, respectively.
lambda tuning parameter

References

Guo W, Lin S. 2008. Generalized linear modeling with regularization for detecting common disease rare haplotype association.


[Package rGLM version 2.0 Index]