hypercube: Diffusion kernels on bi-allelic genotypes

In dkDNA: Diffusion Kernels on a Set of Genotypes

Description

This function construct a diffusion kernel on a p-dimensional hypercube, where each genotype takes on two possible configurations. This graph is obained by the p-Cartesian graph product of a complete graph K_2. It contains 2^p vertices corresponding to sequences of genotypes, and two vertices are adjacent if and only if just one SNP locus differs.

Usage

hypercube(X, theta)

Arguments

X	A genotype matrix of n individuals with p bi-allelic genotypes (n \times p).
theta	The rate of diffusion.

Value

Diffusion kernel matrix of size n \times n. This can be viewed as a covariance among individuals given the diffusion rate.

Author(s)

Gota Morota [e-mail] and Masanori Koyama [e-mail]

References

Kondor R and Lafferty J: (2002) Diffusion Kernels on Graphs and Other Discrete Input Spaces. ICML.

Morota G, Koyama M, Rosa GJM, Weigel KA, and Gianola D. (2013). Predicting complex traits using a diffusion kernel on genetic markers with an application to dairy cattle and wheat data. Genetics Selection Evolution. 45:17.

Examples

# set a seed
set.seed(4321)

# create a genotype matrix of 5 individuals with 10 bi-allelic genotypes
X <- matrix(sample(c(0,1), 50, prob=c(0.6,0.4), replace=TRUE), ncol=10)

# set the rate of diffusion equal to 1
theta <- 1

# compute a hypercube kernel 
hypercube(X, theta)

dkDNA documentation built on May 2, 2019, 5:15 a.m.