Description Usage Arguments Value Author(s) See Also Examples
matrixEpistasis uses large matrix operation to perform the exhaustive epistasis scan for quantitative traits with covariate adjustment
1 | matrixEpistasis(snpA, snpB, trait, covariate = NULL)
|
snpA |
a matrix of numeric values in size of sample*snp representing the 1st group of SNPs, where the column names are the snp_ids |
snpB |
a matrix of numeric values in size of sample*snp representing the 2nd group of SNPs, where the column names are the snp_ids |
trait |
a vector of numeric values representing the quantitative trait |
covariate |
a matrix of numeric values in size of sample*covariate, by default, NULL |
A list containing the follow components:
r a matrix of numeric values representing the partial correlation coefficients between snpA*snpB and traits conditioned on snpA, snpB and covariates
df an integer value representing the degree of freedom
Shijia Zhu, shijia.zhu@mssm.edu
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 | # randomly generate a SNP matrix
snp <- sapply(1:100,function(i) rnorm(1000) )
# assign names to SNPs
colnames(snp) <- paste0('snp',1:100)
snpA = snp
snpB = snp
# radnomly generate a quantitative trait by simulating the relationship between SNPs and traits
trait <- snp %*% rnorm(100)
# use the top 5 PCs as the covariates
covariate <- prcomp(snp)$x[,1:5]
# run matrixEpistasis with covariates adjustment
res <- matrixEpistasis( snpA=snpA, snpB=snpB, trait=trait, covariate = covariate )
r <- res$r
df <- res$df
# run matrixEpistasis function with covariates adjustment
res <- matrixEpistasis( snpA=snpA, snpB=snpB, trait=trait, covariate = covariate )
# res is a list comprising two components: r and df. res$r is the matrix of partial correlation coefficients between snp interaction (snpA*snpB) and trait with additive effects (snpA and snpB) and covariates adjusted, and res$df is the degree of freedom for the partial correlation.
names(res)
r = res$r
df = res$df
# based on the degree of freedom, run matrixPval function to calculate p values for all partial correlation coefficients. The result is a matrix of p-values for epistasis of all snp interactions.
p <- matrixPval( r , df )
# alternatively, users can calculate p-values only for those entries with p vlaues less than a given threshold, say 1e-2, shown as follows:
# use p2c to covert p value threshold 1e-5 to the corresponding partial correlation coefficient
corrThreshold <- p2c( pval=1e-2 , df )
# extract the index for those significant ones
index <- which( abs(r)>corrThreshold , arr.ind=TRUE )
# get the SNP names
snp1 <- colnames(snpA)[ index[,1] ]
snp2 <- colnames(snpB)[ index[,2] ]
# use matrixPval function to calculate p values for only those of interest
pvalue <- matrixPval( r[index] , df )
# build the data frame
sig_res <- data.frame( snp1 , snp2 , pvalue )
head(sig_res)
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.