The aim of this package is to propose several methods for testing gene-gene interaction in case-control association studies. Such a test can be done by aggregating SNP-SNP interaction tests performed at the SNP level (SSI) or by using gene-gene multidimensionnal methods (GGI) methods. The package also proposes tools for a graphic display of the results.


To install and load the package in R


A detailed example

Importation of genotypes with ImportFile function Supported format are pedfile, PLINK, VCF (4.0) file, or genotypes imputed by IMPUTE2.

#### Example of ped format with 17 genes
ped <- system.file("extdata/example.ped", package="GeneGeneInteR")
info <- system.file("extdata/", package="GeneGeneInteR")
posi <- system.file("extdata/example.txt", package="GeneGeneInteR")
dta <- importFile(file=ped, snps=info, pos=posi, pos.sep="\t")

## Importation of the phenotype
resp <- system.file("extdata/response.txt", package="GeneGeneInteR")
Y  <- read.csv(resp, header=FALSE)

Prior to the statistical analysis, dataset can be modified by applying filters to the SNPs (snpMatrixScour function) or by imputing missing genotypes (imputeSnpMatrix function). A subset of genes can also be selected with the select.snps function.

## Filtering of the data: SNPs with MAF < 0.05 or p.value for HWE < 1e-3 are removed. No filtering is applied regarding missing data (call.rate=1).
dta <- snpMatrixScour(dta$snpX, = dta$, min.maf = 0.05, min.eq = 1e-3, call.rate = 1)
## Imputation of the missing genotypes
dta <- imputeSnpMatrix(dta$snpX,dta$
## Selection of a subset of 12 genes
dta <- selectSnps(dta$snpX, dta$, c("bub3","CDSN","Gc","GLRX","PADI1","PADI2","PADI4","PADI6","PRKD3","PSORS1C1","SERPINA1","SORBS1"))

Gene-based gene-gene interaction analysis can be performed by testing each pair of genes in the datatset (function GGI). 10 methods are implemented in the GeneGeneInteR package to test a pair of genes: - 6 Gene-Gene multidimensional methods - Principal Components Analysis - PCA - Canonical Correlation Analysis - CCA - Kernel Canonical Correlation Analysis - KCCA - Composite Linkage Disequilibrium - CLD - Partial Least Square Path Modeling - PLSPM - Gene-Based Information Gain Method - GBIGM - 4 Gene-Gene interaction methods based on SNP-SNP interaction testing: - Minimum p-value test - minP - Gene Association Test using Extended Simes procedure - GATES - Truncated Tail Strength test - tTS - Truncated p-value Product test - tProd .

## Testing for all pair of genes with the CLD method
GGI.res <- GGI(Y=Y, snpX=dta$snpX,$,method="CLD")

Visualization of the results can be performed through either a matrix display (GGI.plot) or a network output (

## Plot of the results with default values
## Plot of the results with a threshold and an ordering of the genes.

## Example of network with default threshold 0.05
## Example of network with threshold 0.01 where genes with no interaction are not plotted (plot.nointer=FALSE)

MathieuEmily/GeneGeneInteR documentation built on Jan. 13, 2018, 6:55 a.m.