episcan {episcan}

Scan pairwise epistasis

Description

Genomic interaction analysis with EPIBLASTER or epistasis-oriented Hilbert–Schmidt Independence Criterion (HSIC).

Usage

episcan(geno1, geno2 = NULL, pheno = NULL,
  phetype = c("case-control", "quantitative"), outfile = "episcan",
  suffix = ".txt", zpthres = 1e-06, chunksize = 1000, scale = TRUE,
  ...)

Arguments

geno1

a data.frame or matrix of the first genotype data. big.matrix object from bigmemory also works. The columns contain the information of variables and the rows contain the information of samples.

geno2

optional. A data.frame or matrix of the second genotype data. big.matrix object from bigmemory also works. The columns contain the information of variables and the rows contain the information of samples.

pheno

a vector (named or not). If not provided, the value of geno2 will be used if it is a vector. The values is either case-control phenotype (0, 1) or quantitative phenotype.

phetype

character string. Either "case-control" or "quantitative".

outfile

output file name. Default is "episcan".

suffix

suffix for output file. Default is ".txt". The final result will be stored in outfilesuffix.

zpthres

is the significance threshold to select variant pairs for output. Default is 1e-6.

chunksize

the number of variants in each chunk.

scale

a logical value to define wheter the input data needs to be normalized. Default is TRUE which means, by default, all the genotype data will be normalized and if the phetype is "quantitative", the phenotype will also be normalized.

...

not used.

Value

null

Author(s)

Beibei Jiang [e-mail]

References

Kam-Thong, T., D. Czamara, K. Tsuda, K. Borgwardt, C. M. Lewis, A. Erhardt-Lehmann, B. Hemmer, et al. 2011. "EPIBLASTER-Fast Exhaustive Two-Locus Epistasis Detection Strategy Using Graphical Processing Units." Journal Article. European Journal of Human Genetics 19 (4): 465–71. https://doi.org/10.1038/ejhg.2010.196.

Kam-Thong, T., B. Pütz, N. Karbalai, B. Müller-Myhsok, and K. Borgwardt. 2011. "Epistasis Detection on Quantitative Phenotypes by Exhaustive Enumeration Using GPUs." Journal Article. Bioinformatics 27 (13): i214–21. https://doi.org/10.1093/bioinformatics/btr218.

Examples

# simulate some data
set.seed(123)
geno1 <- matrix(sample(0:2, size = 1000, replace = TRUE, prob = c(0.5, 0.3, 0.2)), 
ncol = 10)
geno2 <- matrix(sample(0:2, size = 2000, replace = TRUE, prob = c(0.4, 0.3, 0.3)), 
ncol = 20)
dimnames(geno1) <- list(row = paste0("IND", 1:nrow(geno1)), 
col = paste0("rs", 1:ncol(geno1)))
dimnames(geno2) <- list(row = paste0("IND", 1:nrow(geno2)), 
col = paste0("exm", 1:ncol(geno2)))
p1 <- c(rep(0, 60), rep(1, 40))
p2 <- rnorm(100)

# one genotype with case-control phenotype
episcan(geno1 = geno1, 
geno2 = NULL, 
pheno = p1, 
phetype = "case-control",
outfile = "episcan_1geno_cc", 
suffix = ".txt", 
zpthres = 0.9, 
chunksize = 10, 
scale = TRUE)

# take a look at the result
res <- read.table("episcan_1geno_cc.txt", 
header = TRUE, 
stringsAsFactors = FALSE)
head(res)

# two genotypes with quantitative phenotype
episcan(geno1 = geno1, 
geno2 = geno2, 
pheno = p2, 
phetype = "quantitative",
outfile = "episcan_2geno_quant", 
suffix = ".txt", 
zpthres = 0.9, 
chunksize = 10, 
scale = TRUE)