Description Usage Arguments Details Value Author(s) References See Also Examples
Performs a genotypic TDT for geneenvironment interactions for each SNP represented by a column of a matrix in genotype format
and a binary environmental factor. If alpha1
is set to a value smaller than 1, then the twostep procedure of
Gauderman et al. (2010) will be used to first select all SNPs showing a pvalue smaller than alpha1
in a logistic regression of the environmental factor against the sums of the codings for the parents' genotypes at the respective
SNP. In the second step, the genotypic TDT is then applied to the selected SNPs.
If unstructured = TRUE
, all fully parameterized model is considered and a likelihood ratio test is performed.
While colGxE
computes the pvalues based on asymptotic ChiSquaredistributions,
colGxEPerms
can be used to determine permutationbased pvalues for the basic genotypic TDT (i.e. for colGxE
using alpha = 1
and unstructured = FALSE
.
1 2 3 4 5 6  colGxE(mat.snp, env, model = c("additive", "dominant", "recessive"),
alpha1 = 1, size = 50, addGandE = TRUE, whichLRT = c("both", "2df", "1df", "none"),
add2df = TRUE, addCov = FALSE, famid = NULL, unstructured = FALSE)
colGxEPerms(mat.snp, env, model = c("additive", "dominant", "recessive"),
B = 10000, size = 20, addPerms = TRUE, famid = NULL, rand = NA)

mat.snp 
a numeric matrix in which each column represents a SNP. Each column must be
a numeric vector of length 3 * t representing a SNP genotyped at t trios. Each of the t
blocks must consist of the genotypes of father, mother, and offspring
(in this order). The genotypes must be coded by 0, 1, and 2. Missing values are allowed and need to be coded by 
env 
a vector of length t (see 
model 
type of model that should be fitted. Abbreviations are allowed. Thus, e.g., 
alpha1 
a numeric value between 0 and 1 (excluding 0). If 
size 
the number of SNPs considered simultaneously when computing the parameter estimates. 
addGandE 
should the relative risks and their confidence intervals for the exposed cases be added to the output? 
whichLRT 
character string specifying which likelihood ratio test should be added to the output. If 
add2df 
should the results of a 2 df Wald test for testing both the SNP and the interaction effect simultaneously be added to the model? 
addCov 
should the covariance between the parameter estimations for the SNP and the geneenvironment interaction be added
to the output? Default is 
famid 
a vector of the same length as 
unstructured 
should a fully parameterized model be fitted? If 
B 
number of permutations. 
addPerms 
should the matrices containing the permuted values of the test statistics for the SNP and the geneenvironment interaction be added to the output? 
rand 
integer for setting the random number generator into a reproducible state. 
A conditional logistic regression model including two parameters, one for G, and the other for GxE, is fitted, where
G is specified according to model
.
For colGxE
with unstructured=FALSE
, an object of class colGxE
consisting of the following numeric matrices with two columns (one for each parameter):
coef 
the estimated parameter, 
se 
the estimated standard deviation of the parameter estimate, 
stat 
Wald statistic, 
RR 
the relative risk, i.e.\ in the case of trio data, 
lowerRR 
the lower bound of the 95% confidence interval for 
upperRR 
the upper bound of the 95% confidence interval for 
usedTrios 
the number of trios affecting the parameter estimation, 
env 
vector containing the values of the environmental factor, 
type 

addGandE 
the value of 
addOther 
a logical vector specifying which of the likelihood ratio tests and if the 2 df Wald test was performed, 
and depending on the specifications in colGxE
cov 
numeric vector containing the covariances, 
lrt2df 
a numeric matrix with two columns, in which the first column contains the values of the 1 df likelihood ratio test statistic and the second the corresponding pvalues, 
wald2df 
a numeric matrix with two columns, in which the first column contains the values of the 2 df Wald test statistics and the second the corresponding pvalues, 
lrt1df 
a numeric matrix with two columns, in which the first column contains the values of the 2 df likelihood ratio test statistic and the seocnd the corresponding pvalues. 
For colGxE
with unstructured=TRUE
, an object of class colGxEunstruct
consisting of the following vectors:
ll.main 
the loglikelihoods of the models containing only the two main effects, 
ll.full 
the loglikelihoods of the models additionally containing the two main effects and the two interaction effects, 
stat 
the values of the test statistic of the likelihood ratio test, 
pval 
the corresponding pvalues. 
For colGxEPerms
,
stat 
a matrix with two columns containing the values of gTDT statistics for the main effects of the SNPs and the geneenvironment interactions when considering the original, unpermuted casepseudocontrol status, 
pval 
a matrix with two columns comprising the permutationbased pvalues corresponding to the test statistics in 
and if addPerms = TRUE
matPermG 
a matrix with 
matPermGxE 
a matrix with 
Holger Schwender, holger.schwender@udo.edu
Gauderman, W.J., Thomas, D.C., Murcray, C.E., Conti, D., Li, D., and Lewinger, J.P. (2010). Efficient GenomeWide Association Testing of GeneEnvironment Interaction in CaseParent Trios. American Journal of Epidemiology, 172, 116122.
Schaid, D.J. (1996). General Score Tests for Associations of Genetic Markers with Disease Using Cases and Their Parents. Genetic Epidemiology, 13, 423449.
Schwender, H., Taub, M.A., Beaty, T.H., Marazita, M.L., and Ruczinski, I. (2011). Rapid Testing of SNPs and GeneEnvironment Interactions in CaseParent Trio Data Based on Exact Analytic Parameter Estimation. Biometrics, 68, 766773.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16  # Load the simulated data for the analysis.
data(trio.data)
# Set up a vector with the binary environmental variable.
# Here, we consider the genegender interactions and
# assume that the children in the first 50 trios are
# girls, and the remaining 50 are boys.
sex < rep(0:1, each = 50)
# Test the interaction of sex with each of the SNPs in mat.test
gxe.out < colGxE(mat.test, sex)
# By default, an additive mode of inheritance is considered.
# If, e.g., a dominant mode should be considered, then this can
# be done by calling
gxeDom.out < colGxE(mat.test, sex, model="dominant")

Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.