| Title: | Epistasis Analysis for Quantitative Traits by Functional Regression Model |
|---|---|
| Description: | A Tool for Epistasis Analysis Based on Functional Regression Model |
| Authors: | Futao Zhang |
| Maintainer: | Futao Zhang <[email protected]> |
| License: | GPL-2 |
| Version: | 1.43.0 |
| Built: | 2024-10-30 07:16:55 UTC |
| Source: | https://github.com/bioc/FRGEpistasis |
FRGEpistasis is designed to detect the epistasis between genes or genomic regions for both common variants and rare variants. Currently FRGEpistasis was developed by Futao Zhang with R language and maintained in Xiong lab at UTSPH. This tool is friendly, convenient and memory efficient.
| Package: | FRGEpistasis |
| Type: | Package |
| Version: | 0.99.5 |
| Date: | 2014-03-22 |
| License: | GPL-2 |
It currently has the following functional modules: Functional Regression Model (FRG) for Testing Interaction; Regression on Principal Components Analysis (PCA) for Testing Interaction; Point-wise interaction Test.
Futao Zhang Maintainer: Futao Zhang <[email protected]>
This function aims to expand the genotype of one gene (or genomic region) with Fourier Expansion.
fourierExpansion(gene_idx, geno, gene_list, snp_map, rng)fourierExpansion(gene_idx, geno, gene_list, snp_map, rng)
gene_idx |
The expansion gene index in the gene annotation list. |
geno |
Genotype of all the genes in the gene annotation list. |
gene_list |
Gene annotation list which includes gene name, chromosome, start position and end position. |
snp_map |
SNP genetic map includes chromosome, snp indentifier, genetic distance and base-pair position. |
rng |
A numeric value which represents gene region extensible scope. |
This function reduces the dimension of one gene(or genomic region) with Fourier Expansion. Fist extract out the genotype of this gene with the gene annotation and the SNP map information. Then expanse the gene with the genotypes and SNP positions if the number of SNPs in the gene is over 3. Otherwise the raw genotypes of the gene would be returned. The number of Fourier Basis is selected to explain 80 percent of genetic variation.
If the SNPs number of the gene is over 3, returns the expansion of the genotype which is a matrix with the dimension Sample number * Fourier Basis number. If the SNPs number of the gene is no more than 3, returns the raw genotype of the gene which is a matrix with the dimension Sample number * SNP number.
Futao Zhang
gLst<-read.csv(system.file("extdata", "gene.list.csv", package="FRGEpistasis")) fdata<-read.table(system.file("extdata", "simGeno-chr1.raw", package="FRGEpistasis"),header=TRUE) geno<-fdata[,-1:-6] snp_map<-read.table(system.file("extdata", "chr1.map", package="FRGEpistasis")) fourierExpansion(1, geno, gLst, snp_map, 0)gLst<-read.csv(system.file("extdata", "gene.list.csv", package="FRGEpistasis")) fdata<-read.table(system.file("extdata", "simGeno-chr1.raw", package="FRGEpistasis"),header=TRUE) geno<-fdata[,-1:-6] snp_map<-read.table(system.file("extdata", "chr1.map", package="FRGEpistasis")) fourierExpansion(1, geno, gLst, snp_map, 0)
This function is the entrance of the software package. It tests the Genome Wide Epistasis by Functional Regression Model.
fRGEpistasis(wDir, phenoInfo, gnoFiles, mapFiles, gLst, fdr, rng)fRGEpistasis(wDir, phenoInfo, gnoFiles, mapFiles, gLst, fdr, rng)
wDir |
The dataset directory. If the dataset is in the working directory, wDir is ".". |
phenoInfo |
It is a matrix with two columns. One column is the individual ID and the other is the phenotype. The phenotype can be quantitative trait or binary trait. |
gnoFiles |
The vector of genotype file names. It contains the genotype file names indicating where to read the genotype files. |
mapFiles |
The vector of SNP genetic map file names. It contains the map file names indicating where to read the genetic map files. |
gLst |
Gene annotation which includes gene name, chromosome, start position and end position. |
fdr |
FDR control threshold, When this value == 1, turn FDR control off. |
rng |
A numeric value which represents gene region extensible scope. |
Firstly this package reduces the dimension of genotype of all the genomic regions. Secondly this package tests the epistasis of genomic regions both of which are on the same chromosome(file). Thirdly this package tests the epistasis of genomic regions which are on different chromosomes(files).
This function is memory efficient with high performance. Memory efficiency: Only store reduced expansion data of genotypes instead of raw data of genotypes. This package reduces the dimension of genotype of all the genomic regions(see details of function "reduceGeno"). In real dataset the genotypes on different chromosome are always organized into different files. And each genotype file is very large. Reading all the files into memory is unacceptable. This package reads the files one by one and reduces the genotype dimension with Fourier expansion. In order to inform the package how many files and where to read, we need two data structures "gnoFiles" and "mapFiles" to store the file names.
high performance: Each data file only needs to read once and reduce dimension once. So I/O times are reduced and repeated computing of data reduction was avoided. This method is a kind of group test. We take a gene(or genomic region) as the test unit. The number of Test is sharply reduced comparing with point-wise interaction (SNP-SNP) test. The dimension of genotype is reduced by functional expansion, So the time of each test is reduced.
Return a data frame which contains all the names of gene pairs and the p values of chi-square test for their epistasis.
Futao Zhang
wDir <-paste(system.file("extdata", package="FRGEpistasis"),"/",sep="") gnoFiles<-read.table(system.file("extdata", "list_geno.txt", package="FRGEpistasis")) mapFiles<-read.table(system.file("extdata", "list_map.txt", package="FRGEpistasis")) phenoInfo <- read.csv(system.file("extdata", "phenotype.csv", package="FRGEpistasis"),header=TRUE) gLst<-read.csv(system.file("extdata", "gene.list.csv", package="FRGEpistasis")) rng=0 fdr=0.05 out_epi <- data.frame( ) phenoInfo [,2]=log(phenoInfo [,2]) out_epi = fRGEpistasis(wDir,phenoInfo,gnoFiles,mapFiles,gLst,fdr,rng)wDir <-paste(system.file("extdata", package="FRGEpistasis"),"/",sep="") gnoFiles<-read.table(system.file("extdata", "list_geno.txt", package="FRGEpistasis")) mapFiles<-read.table(system.file("extdata", "list_map.txt", package="FRGEpistasis")) phenoInfo <- read.csv(system.file("extdata", "phenotype.csv", package="FRGEpistasis"),header=TRUE) gLst<-read.csv(system.file("extdata", "gene.list.csv", package="FRGEpistasis")) rng=0 fdr=0.05 out_epi <- data.frame( ) phenoInfo [,2]=log(phenoInfo [,2]) out_epi = fRGEpistasis(wDir,phenoInfo,gnoFiles,mapFiles,gLst,fdr,rng)
This function is used to analyse the epistasis of genomic region A and genomic region B.
frgEpistasisTest(pheno, geno_A, pos_A, geno_B, pos_B)frgEpistasisTest(pheno, geno_A, pos_A, geno_B, pos_B)
pheno |
A vector of phenotype which can be quantitative trait or binary trait. |
geno_A |
Genotype matrix of gene ( or genomic region) A. |
pos_A |
Vector of physical positions of SNPs in gene ( or genomic region) A. |
geno_B |
Genotype matrix of gene ( or genomic region) B. |
pos_B |
Vector of physical positions of SNPs in gene ( or genomic region) B. |
This function is independent with other functions in this package. It is designed for small dataset test. It takes phenotype, genotype and Physical positions as the input. If the position information is NULL, this function considers the SNPs in this gene to be uniformly filled in the gene scope. First this function expanses the genotypes of gene A and gene B. Then it analyses their epistasis.
It returns the p value of chi-square test for epistasis detection between gene A and gene B.
Futao Zhang
smp_num=1000 number_snp_A=25 number_snp_B=20 pheno<-sample(c(0:500),smp_num,replace=TRUE) smpl=rep(0,number_snp_A*smp_num) idx_1=sample(c(1:(number_snp_A*smp_num)),ceiling(number_snp_A*smp_num/100)) idx_2=sample(c(1:(number_snp_A*smp_num)),ceiling(number_snp_A*smp_num/200)) smpl[idx_1]=1 smpl[idx_2]=2 geno_A=matrix(smpl,smp_num,number_snp_A) smpl=sample(c(0,1,2),number_snp_B*smp_num,replace=TRUE) geno_B=matrix(smpl,smp_num,number_snp_B) frgEpistasisTest(pheno,geno_A,pos_A=NULL,geno_B,pos_B=NULL)smp_num=1000 number_snp_A=25 number_snp_B=20 pheno<-sample(c(0:500),smp_num,replace=TRUE) smpl=rep(0,number_snp_A*smp_num) idx_1=sample(c(1:(number_snp_A*smp_num)),ceiling(number_snp_A*smp_num/100)) idx_2=sample(c(1:(number_snp_A*smp_num)),ceiling(number_snp_A*smp_num/200)) smpl[idx_1]=1 smpl[idx_2]=2 geno_A=matrix(smpl,smp_num,number_snp_A) smpl=sample(c(0,1,2),number_snp_B*smp_num,replace=TRUE) geno_B=matrix(smpl,smp_num,number_snp_B) frgEpistasisTest(pheno,geno_A,pos_A=NULL,geno_B,pos_B=NULL)
Test interaction between two gene (or genomic regions) with chi-squared test.
fRGInteraction(phenoData, x_A, x_B)fRGInteraction(phenoData, x_A, x_B)
phenoData |
Vector of phenotype data which can be quantitative trait or binary trait. |
x_A |
Expansion data matrix of Genotype of gene A. |
x_B |
Expansion data matrix of Genotype of gene B. |
This function takes phenotype vector and expansed genotype matrices as input. It is the most important part of this software package. It is called by functions "innerEpi" and "innerEpi" of this package. The interaction between gene A and gene B is tested with chi-squared test.
It returns the p value of chi-squared test for epistasis detection between gene A and gene B.
Futao Zhang
x_A<-as.matrix(rnorm(1000,mean=0,sd=1)) x_B<-as.matrix(rnorm(1000,mean=0,sd=1)) phenoData<-runif(1000,15,60) fRGInteraction(phenoData,x_A,x_B)x_A<-as.matrix(rnorm(1000,mean=0,sd=1)) x_B<-as.matrix(rnorm(1000,mean=0,sd=1)) phenoData<-runif(1000,15,60) fRGInteraction(phenoData,x_A,x_B)
Detect epistasis between 2 genes (or genomic regions) both of which are on the same chromosome.
innerEpi(pheno, gstnd, geno_expn, gname, gchr)innerEpi(pheno, gstnd, geno_expn, gname, gchr)
pheno |
Vector of phenotype data which can be quantitative trait or binary trait. |
gstnd |
Vector of indexes which indicates the start indexes and end indexes of expansed genotype of each gene on current chromosome in the matrix "geno_expn". |
geno_expn |
Matrix of expansed genotype data of all the genes. |
gname |
Vector of gene names on current chromosome. |
gchr |
Vector of Chromosome number of current chromosome. |
This function tests the epistasis between 2 genes both of which are on the same chromosome. It takes expansed genotype data as the input. First the data of the gene are extracted from "geno_expn" with "gstnd" and "gname". Then the function "fRGInteraction" will be called.
Return a matrix which contains the gene names of the gene pairs and the p values of chi-squared test for the epistasis of the gene pairs.
Futao Zhang
smp_num<-1000 number_basis<-9 pheno<-sample(c(0:500),smp_num,replace=TRUE) gname<-c("g1","g2") gstnd<-c(0,5,9) smpl<-runif(number_basis*smp_num, 0.0, 1.0) geno_expn<-matrix(smpl,smp_num,number_basis) gchr<-c(1,1) innerEpi(pheno,gstnd,geno_expn,gname,gchr)smp_num<-1000 number_basis<-9 pheno<-sample(c(0:500),smp_num,replace=TRUE) gname<-c("g1","g2") gstnd<-c(0,5,9) smpl<-runif(number_basis*smp_num, 0.0, 1.0) geno_expn<-matrix(smpl,smp_num,number_basis) gchr<-c(1,1) innerEpi(pheno,gstnd,geno_expn,gname,gchr)
Test the SNP-SNP interaction. And the SNPs are organized into one data structure.
innerSnpListInteraction(pheno, snpList)innerSnpListInteraction(pheno, snpList)
pheno |
Vector of phenotype data. |
snpList |
Matrix of the genotypes of all the SNPs for testing the pairwise interactions. |
This function aims to test the pairwise interactions between the SNPs organized into the same data structure. It takes phenotype and genotypes of the SNPs as the input. And output all the p values for the interactions of SNP pairs.
Return a frame contains names of all the SNPs pairs and p values for interactions of these pairs.
Futao Zhang
pheno<- round(runif(1000,40,60)) geno<- as.data.frame(matrix(round(runif(5000,0,2)),1000,5)) innerSnpListInteraction(pheno,geno)pheno<- round(runif(1000,40,60)) geno<- as.data.frame(matrix(round(runif(5000,0,2)),1000,5)) innerSnpListInteraction(pheno,geno)
Logarithmic Transformation of Phenotype
logTransPheno(pheno)logTransPheno(pheno)
pheno |
Vector of phenotype which is the quantitative trait. |
Some variables are not normally distributed. And using statistical tests on this data can give misleading results because they do not meet the statistical assumptions. Many variables have log-normal distributions.
Return vector of transformed phenotype.
smp_num=100 pheno<-sample(c(0:500),smp_num,replace=TRUE) logTransPheno(pheno)smp_num=100 pheno<-sample(c(0:500),smp_num,replace=TRUE) logTransPheno(pheno)
Detect epistasis between 2 genes (or genomic regions) which are on different chromosomes.
outerEpi(pheno, gstnd, gStndp, geno_expn, gname, gNamep, gchr, gChrp)outerEpi(pheno, gstnd, gStndp, geno_expn, gname, gNamep, gchr, gChrp)
pheno |
Vector of phenotype data which can be quantitative trait or binary trait. |
gstnd |
Vector of indexes which indicates the start indexes and end indexes of expansed genotype of each gene on one chromosome in the matrix "geno_expn". |
gStndp |
Vector of indexes which indicates the start indexes and end indexes of expansed genotype of each gene on the other chromosome in the matrix "geno_expn". |
geno_expn |
Matrix of expansed genotype data of all the genes. |
gname |
Vector of gene names on one chromosome. |
gNamep |
Vector of gene names on the other chromosome. |
gchr |
Vector of Chromosome number of one chromosome. |
gChrp |
Vector of Chromosome number of the other chromosome. |
This function tests the epistasis between 2 genes which are on different chromosomes. It takes expansed genotype data as the input. First the data of the gene are extracted from "geno_expn" with "gstnd" and "gname". Then the function "fRGInteraction" will be called.
Return a matrix which contains the gene names of the gene pairs and the p values of chi-squared test for the epistasis of the gene pairs.
Futao Zhang
smp_num=1000 number_basis<-40 pheno<-sample(c(0:500),smp_num,replace=TRUE) gname<-c("g1","g2") gNamep<-c("r1","r2","r3") gstnd<-c(0,5,9) gStndp<-c(16,23,29,36) smpl<-runif(number_basis*smp_num, 0.0, 1.0) geno_expn<-matrix(smpl,smp_num,number_basis) gchr<-c(1,1) gchrp<-c(3,3,3) outerEpi(pheno,gstnd,gStndp,geno_expn,gname,gNamep,gchr,gchrp)smp_num=1000 number_basis<-40 pheno<-sample(c(0:500),smp_num,replace=TRUE) gname<-c("g1","g2") gNamep<-c("r1","r2","r3") gstnd<-c(0,5,9) gStndp<-c(16,23,29,36) smpl<-runif(number_basis*smp_num, 0.0, 1.0) geno_expn<-matrix(smpl,smp_num,number_basis) gchr<-c(1,1) gchrp<-c(3,3,3) outerEpi(pheno,gstnd,gStndp,geno_expn,gname,gNamep,gchr,gchrp)
Test the SNP-SNP interaction. And the SNPs are organized into two different SNP Lists.
outerSnpListInteraction(pheno, snpList1, snpList2)outerSnpListInteraction(pheno, snpList1, snpList2)
pheno |
Vector of phenotype data. |
snpList1 |
Matrix of the genotypes of all the SNPs on the first SNP list for testing the pairwise interactions. |
snpList2 |
Matrix of the genotypes of all the SNPs on the second SNP list for testing the pairwise interactions. |
This function aims to test the pairwise interactions between the SNPs organized into different data structures. It takes phenotype and genotypes of the SNPs as the input. And output all the p values for the interactions of SNP pairs.
Return a frame contains names of all the SNPs pairs and p values for interactions of these pairs.
Futao Zhang
snp_list_1 <- as.data.frame(matrix(round(runif(3000,0,2)),1000,3)) snp_list_2 <- as.data.frame(matrix(round(runif(5000,0,2)),1000,5)) colnames(snp_list_1 )<-c("rs10","rs11","rs12") colnames(snp_list_2 )<-c("rs20","rs21","rs22","rs23","rs24") pheno<- round(runif(1000,40,60)) outerSnpListInteraction(pheno,snp_list_1,snp_list_2)snp_list_1 <- as.data.frame(matrix(round(runif(3000,0,2)),1000,3)) snp_list_2 <- as.data.frame(matrix(round(runif(5000,0,2)),1000,5)) colnames(snp_list_1 )<-c("rs10","rs11","rs12") colnames(snp_list_2 )<-c("rs20","rs21","rs22","rs23","rs24") pheno<- round(runif(1000,40,60)) outerSnpListInteraction(pheno,snp_list_1,snp_list_2)
Test the epistasis between two genes (or genomic regions) with the principal components analysis method.
pCAInteraction(phenoData, x_A, x_B)pCAInteraction(phenoData, x_A, x_B)
phenoData |
Vector of phenotype data which can be quantitative trait or binary trait. |
x_A |
Matrix of genotype of gene A. |
x_B |
Matrix of genotype of gene B. |
This function takes phenotype vector and genotype matrices as input and tests the epistasis using PCA method. The number of principal components is determined by PCA to explain 80 percent of the genetic variation. The interaction between gene A and gene B is tested with chi-squared test.
It returns the p value of chi-squared test for epistasis detection between gene A and gene B.
Futao Zhang
smp_num=1000 number_snp_A=25 number_snp_B=20 pheno<-sample(c(0:500),smp_num,replace=TRUE) smpl=rep(0,number_snp_A*smp_num) idx_1=sample(c(1:(number_snp_A*smp_num)),ceiling(number_snp_A*smp_num/100)) idx_2=sample(c(1:(number_snp_A*smp_num)),ceiling(number_snp_A*smp_num/200)) smpl[idx_1]=1 smpl[idx_2]=2 geno_A=matrix(smpl,smp_num,number_snp_A) smpl=sample(c(0,1,2),number_snp_B*smp_num,replace=TRUE) geno_B=matrix(smpl,smp_num,number_snp_B) pCAInteraction(pheno,geno_A,geno_B)smp_num=1000 number_snp_A=25 number_snp_B=20 pheno<-sample(c(0:500),smp_num,replace=TRUE) smpl=rep(0,number_snp_A*smp_num) idx_1=sample(c(1:(number_snp_A*smp_num)),ceiling(number_snp_A*smp_num/100)) idx_2=sample(c(1:(number_snp_A*smp_num)),ceiling(number_snp_A*smp_num/200)) smpl[idx_1]=1 smpl[idx_2]=2 geno_A=matrix(smpl,smp_num,number_snp_A) smpl=sample(c(0,1,2),number_snp_B*smp_num,replace=TRUE) geno_B=matrix(smpl,smp_num,number_snp_B) pCAInteraction(pheno,geno_A,geno_B)
This function is the another entrance of the software package. It tests the Genome Wide Epistasis by PCA Method and Piontwise Method.
pCAPiontwiseEpistasis(wDir, oEpi, phenoInfo, gnoFiles, mapFiles, gLst, rng)pCAPiontwiseEpistasis(wDir, oEpi, phenoInfo, gnoFiles, mapFiles, gLst, rng)
wDir |
The dataset directory. If the dataset is in the working directory, wDir is ".". |
oEpi |
Output data frame which contains all the names of gene pairs and the p values for their epistasis. |
phenoInfo |
It is a matrix with two columns. One column is the individual ID and the other is the phenotype. The phenotype can be quantitative trait or binary trait. |
gnoFiles |
The vector of genotype file names. It contains the genotype file names indicating where to read the genotype files. |
mapFiles |
The vector of SNP genetic map file names. It contains the map file names indicating where to read the genetic map files. |
gLst |
Gene annotation which includes gene name, chromosome, start position and end position. |
rng |
A numeric value which represents gene region extensible scope. |
The genotypes on different chromosome are stored in different files. The full names of these files are listed in the index file that is taken as the input parameter. After the index file is loaded, the function knows where to read the genotypes files. This function analyses the genotypes files one by another. That means this function tests the epistasis of genomic regions both of which are on the same chromosome(file), then the epistasis of genomic regions which are on different chromosomes(files). This function can test epistasis both with PCA method and pointwise method. For a pair of genes, we assume that the total number of all possible SNP pairs is K, The minum p value for SNP-SNP interaction among the K pairs is output as the pointwise method result of the gene pair.
Return a data frame which contains all the names of gene pairs and the p values of chi-square test for their epistasis.
Futao Zhang
work_dir <-paste(system.file("extdata", package="FRGEpistasis"),"/",sep="") ##read the list of genotype files geno_files<-read.table(system.file("extdata", "list_geno.txt", package="FRGEpistasis")) ##read the list of map files mapFiles<-read.table(system.file("extdata", "list_map.txt", package="FRGEpistasis")) ##read the phenotype file phenoInfo <- read.csv(system.file("extdata", "phenotype.csv", package="FRGEpistasis"),header=TRUE) ##read the gene annotation file gLst<-read.csv(system.file("extdata", "gene.list.csv", package="FRGEpistasis")) ##define the extension scope of gene region rng=0 ##log transformation phenoInfo [,2]=log(phenoInfo[,2]) out_epi<-data.frame() pCAPiontwiseEpistasis(work_dir,out_epi,phenoInfo,geno_files,mapFiles,gLst,rng)work_dir <-paste(system.file("extdata", package="FRGEpistasis"),"/",sep="") ##read the list of genotype files geno_files<-read.table(system.file("extdata", "list_geno.txt", package="FRGEpistasis")) ##read the list of map files mapFiles<-read.table(system.file("extdata", "list_map.txt", package="FRGEpistasis")) ##read the phenotype file phenoInfo <- read.csv(system.file("extdata", "phenotype.csv", package="FRGEpistasis"),header=TRUE) ##read the gene annotation file gLst<-read.csv(system.file("extdata", "gene.list.csv", package="FRGEpistasis")) ##define the extension scope of gene region rng=0 ##log transformation phenoInfo [,2]=log(phenoInfo[,2]) out_epi<-data.frame() pCAPiontwiseEpistasis(work_dir,out_epi,phenoInfo,geno_files,mapFiles,gLst,rng)
Test the epistasis of the gene pair by pointwise method
pointwiseInteraction(phenoData, x_A, x_B)pointwiseInteraction(phenoData, x_A, x_B)
phenoData |
Vector of phenotype data which can be quantitative trait or binary trait. |
x_A |
Matrix of genotype of gene A. |
x_B |
Matrix of genotype of gene B. |
This function takes phenotype vector and genotype matrices as input and tests the epistasis using pointwise method. For a pair of genes, we assume that the total number of all possible SNP pairs is K (one SNP from one gene and the other SNP from the other gene). The interaction of each SNP pair between the two genes is tested. The minum p value for SNP-SNP interaction among the K pairs is output as the pointwise method result of the gene pair.
Return the minum p value for SNP-SNP interaction among the K pairs
Futao Zhang
smp_num=1000 number_snp_A=25 number_snp_B=20 pheno<-sample(c(0:500),smp_num,replace=TRUE) smpl=rep(0,number_snp_A*smp_num) idx_1=sample(c(1:(number_snp_A*smp_num)),ceiling(number_snp_A*smp_num/100)) idx_2=sample(c(1:(number_snp_A*smp_num)),ceiling(number_snp_A*smp_num/200)) smpl[idx_1]=1 smpl[idx_2]=2 geno_A=matrix(smpl,smp_num,number_snp_A) smpl=sample(c(0,1,2),number_snp_B*smp_num,replace=TRUE) geno_B=matrix(smpl,smp_num,number_snp_B) pointwiseInteraction(pheno,geno_A,geno_B)smp_num=1000 number_snp_A=25 number_snp_B=20 pheno<-sample(c(0:500),smp_num,replace=TRUE) smpl=rep(0,number_snp_A*smp_num) idx_1=sample(c(1:(number_snp_A*smp_num)),ceiling(number_snp_A*smp_num/100)) idx_2=sample(c(1:(number_snp_A*smp_num)),ceiling(number_snp_A*smp_num/200)) smpl[idx_1]=1 smpl[idx_2]=2 geno_A=matrix(smpl,smp_num,number_snp_A) smpl=sample(c(0,1,2),number_snp_B*smp_num,replace=TRUE) geno_B=matrix(smpl,smp_num,number_snp_B) pointwiseInteraction(pheno,geno_A,geno_B)
Rank-Based Inverse Normal Transformation of Phenotype
rankTransPheno(pheno, para_c)rankTransPheno(pheno, para_c)
pheno |
Vector of phenotype which is the quantitative trait. |
para_c |
Adjust parameter, commonly as 0,1/3,3/8 or 1/2. |
Some variables are not normally distributed. And using statistical tests on this data can give misleading results because they do not meet the statistical assumptions. This function implements Rank-Based Inverse Normal Transformation to make phenotype normally distributed.
Return vector of rank-based inverse normal transformed phenotype.
Futao Zhang
T. Mark Beasley, Stephen Erickson and David B. Allison. Rank-Based Inverse Normal Transformations are Increasingly Used, But are They Merited? Behav Genet. 2009 Sep.;39(5):580-595.
c=0.5 smp_num=100 pheno<-sample(c(0:500),smp_num,replace=TRUE) rankTransPheno(pheno,c)c=0.5 smp_num=100 pheno<-sample(c(0:500),smp_num,replace=TRUE) rankTransPheno(pheno,c)
Reduce Dimension of Genotype using Functional Regression Model
reduceGeno(wDir, pheno, gnoFiles, mapFiles, gLst, rng)reduceGeno(wDir, pheno, gnoFiles, mapFiles, gLst, rng)
wDir |
The dataset directory. If the dataset is in the working directory, wDir is ".". |
pheno |
It is a matrix with two columns. One column is the individual ID and the other is the phenotype. The phenotype can be quantitative trait or binary trait. |
gnoFiles |
The vector of genotype file names. It contains the genotype file names indicating where to read the genotype files. |
mapFiles |
The vector of SNP genetic map file names. It contains the map file names indicating where to read the genetic map files. |
gLst |
Gene annotation which includes gene name, chromosome, start position and end position. |
rng |
A numeric value which represents gene region extensible scope. |
This function reduces the dimension of genotypes of all the genes with Fourier expansion. In real dataset the genotypes on different chromosome are always organized into different files. And each genotype file is very large. This function processes the genotype files in turns. The reduced genotype data (the expansion data) of all the chromosomes are combined together. The expansion data and other information are organized into a list. During Fourier expansion, the physical position information of the SNPs are used. This is one of merits of our method.
Return a list that includes reduced genotype data, gene names,chromosome information, start index and end index of each gene.
Futao Zhang
wDir <-paste(system.file("extdata", package="FRGEpistasis"),"/",sep="") gnoFiles<-read.table(system.file("extdata", "list_geno.txt", package="FRGEpistasis")) mapFiles<-read.table(system.file("extdata", "list_map.txt", package="FRGEpistasis")) phenoInfo <- read.csv(system.file("extdata", "phenotype.csv", package="FRGEpistasis"),header=TRUE) gLst<-read.csv(system.file("extdata", "gene.list.csv", package="FRGEpistasis")) rng=0 gInfo=reduceGeno(wDir,phenoInfo,gnoFiles,mapFiles,gLst,rng)wDir <-paste(system.file("extdata", package="FRGEpistasis"),"/",sep="") gnoFiles<-read.table(system.file("extdata", "list_geno.txt", package="FRGEpistasis")) mapFiles<-read.table(system.file("extdata", "list_map.txt", package="FRGEpistasis")) phenoInfo <- read.csv(system.file("extdata", "phenotype.csv", package="FRGEpistasis"),header=TRUE) gLst<-read.csv(system.file("extdata", "gene.list.csv", package="FRGEpistasis")) rng=0 gInfo=reduceGeno(wDir,phenoInfo,gnoFiles,mapFiles,gLst,rng)
Test the interaction of one SNP with another
snpPairInteraction(pheno, snp1, snp2)snpPairInteraction(pheno, snp1, snp2)
pheno |
Vector of phenotype data which can be quantitative trait or binary trait. |
snp1 |
Vector of genotype data of SNP1. |
snp2 |
Vector of genotype data of SNP2. |
This function tests the interaction of one SNP with another.
Return the p value for snp-snp interaction
Futao Zhang
pheno<- round(runif(1000,40,60)) snp1<-round(runif(1000,0,2)) snp2<-round(runif(1000,0,2)) pval=snpPairInteraction(pheno,snp1,snp2)pheno<- round(runif(1000,40,60)) snp1<-round(runif(1000,0,2)) snp2<-round(runif(1000,0,2)) pval=snpPairInteraction(pheno,snp1,snp2)