W-test: dataset adaptive association test for main and interaction effects in GWAS data

  JC School of Public Health and Primary Care (JC SPHPC)  |  Faculty of Medicine |  The Chinese University of Hong Kong (CUHK)  |  Hong Kong SAR

Method

Download

Guide
      - R

      - C++

      - Linux

CCRB-Statgene

  Guide for R package

     Input file format:

   genotype.txt (in rows: subjects, in columns: coded SNPs; N x P matrix, N=1,000, P=50); 

   phenotype.txt , (N x 2 matrix, the 1st column is sample ID, and the 2nd column is phenotype).

     Read in data:

  install.packages("wtest")

  library(wtest)

  geno<- read.table(file="genotype.txt", header=TRUE)

  phen<- unlist(read.table(file="phenotype.txt", header=FALSE, row.names=1))

    #  Calculate main effect or pairwise interaction test: using default h and f parameters:

   #input y as a numeric vector, data as N by P matrix.

    w.out1<- wtest(data=geno, y=phen, w.order = 1)

  Step 1.  Estimate h and f parameters: Using bootstrapped samples under null hypothesis to estimate h and f. The option in the function hf.calculation includes n.marker and n.sample. We suggest using all the samples in the data to perform the estimation, which is the default option. When sample size is smaller than 1,000, B should be increased to 400, or 1000, and n.marker is suggested to use 100 or above.

    #Estimate h and f for main effect

    hf1<-hf.calculation(data = geno, B=400, w.order = 1, n.marker=1000, n.sample=1000)

  #Estimate h and f for interaction effect

  hf2<-hf.calculation(data = geno, B=400, w.order = 2, n.marker=50, n.sample=1000)

Step 2A. Calculate pairwise interactions exhaustively using estimated h and f.

  w.out2<-wtest(data = geno, y = phen, w.order=2, input.pval= NULL, hf2 = hf2)

  w.pairs<- w.out2$results

Step 2B. Calculate pairwise interactions using selected main effect markers:

 #Input.pval = 0.1:  only input SNPs which main effect p-values are smaller than 0.1 

 #Output.pval = 0.0005:  output only the pairs that p-values are smaller than 0.0005.

    w.out3<-wtest(data = geno, y = phen, w.order=2, hf1=hf1, hf2 = hf2, input.pval= 0.1, output.pval =0.0005)

  w.pairs<- w.out3$results

#output pairs:

write.table(w.pairs, file= "w.pairs.txt", quote=F, col.names=T, row.names=F, sep= "\t")

Step 3A. Distribution plot:  Diagnostic checking on estimated probability distributions with hypothesis probability distributions. The function will output two distribution curves in the same plot, for each k (number of categories formed by the biomarker), or h and f. 

w.diagnosis(data=geno, w.order = 2, n.rep = 100, hf2=hf2, main=NULL, xlab=NULL, ylab=NULL)

Step 3B. qq plot:  Diagnostic checking on whether there is inflation of type I error

w.qqplot(geno, w.order = 2, hf2= hf2)

abline(0,1)

---- End of the R package tutorial -----

  If you find any problems of the program, please drop a message to: maggiew@cuhk.edu.hk

Copyright