rrBLUP (version 3.2)

impute: Nearest-neighbor imputation

Description

Imputes alleles based on the nearest-neighbor average.

Usage

impute(G, n.neighbor = "ALL", D = NULL)

Arguments

G

Matrix ($n \times m$) of unphased genotypes for $n$ lines and $m$ markers.

n.neighbor

Number of nearest neighbors to use for imputation. If not specified, the population mean is used.

D

Distance matrix ($n \times n$) between lines. If not specified, calculated by dist.

Value

• The imputed G matrix.

Details

Missing data in neighbors are omitted for the average. If all of the neighbors are missing at any marker, an error is returned. This function assumes markers are coded numerically, e.g., {aa,Aa,AA} = {-1,0,1} for bi-allelic markers. Imputed alleles are fractional numbers.

Examples

#random population of 200 lines with 1000 markers
#5% missing data
G <- matrix(rep(0,200*1000),200,1000)
for (i in 1:200) {
  for (j in 1:1000) {
    G[i,j] <- ifelse(runif(1)<0.05,NA,ifelse(runif(1)<0.5,-1,1))
  }
}
G.impute <- impute(G,n.neighbor=20)