OmniGGInterTest: Omnibus gene-gene interaction test

In fhebert/GeneGeneInteractions2: Global tests for the detection of gene-gene interaction effects on binary phenotypes

Description

OmniGGInterTest can be used to perform an omnibus gene-gene interaction test. Given two genes and a binary phenotypes, two interaction models are constructed (one considering a continuous modelization of the SNPs and another considering a dummy coding) and corresponding test statistics vectors are computed. On each vector, the minP, higher criticism, L2-norm and Hotelling methods are applied. Finally, an overall p-value is computed by combining the 8 tests using Simes' and Fisher's methods and truncated versions of these methods.

Usage

OmniGGInterTest(X1, X2, Y, N = 1000, perm.method = "parametric", 
                omni.method = c("simes", "tsimes", "fisher", 
                "tfisher", "simes.MGFR", "fisher.MGFR"))

Arguments

X1	a n x p1 matrix of genotypes (coded as 0/1/2)
X2	a n x p2 matrix of genotypes (coded as 0/1/2)
Y	a vector of length n containing the values of the binary phenotype (0/1)
N	the desired number of permutations of the phenotype (default = 1000)
perm.method	a character giving the chosen resampling method for the phenotype: paramatric bootstrap (perm.method="parametric") or simple permutations (any other character)
omni.method	a vector giving the methods to use for the omnibus combination, among "simes", "tsimes", "fisher", "tfisher", "simes.MGFR" and "fisher.MGFR". By default, the six methods are used

Value

A list containing the following elements:

res_omni	the vector of omnibus p-values
res_continuous	the p-values of the minP, higher criticism and L2-norm methods using the continuous coding
res_dummy	the p-values of the minP, higher criticism and L2-norm methods using the dummy coding
pval0_continuous	the resampling-based p-values of the minP, higher criticism and L2-norm methods using the continuous coding
pval0_dummy	the resampling-based p-values of the minP, higher criticism and L2-norm methods using the dummy coding
fit_continuous	the vector of test statistics, corresponding correlation matrix and vectors of test statistics obtained on permutations using the continuous coding
fit_dummy	the vector of test statistics, corresponding correlation matrix and vectors of test statistics obtained on permutations using the dummy coding

fhebert/GeneGeneInteractions2 documentation built on May 19, 2019, 12:36 a.m.