calc_IBD: Calculation of IBD

In Alethere/SmoothDescent: Application of map-based genotype correction algorithm Smooth Descent

Calculation of IBD

Description

This function computes Identity-by-descent probabilities according to a different IBD models: - Naive model based on random assortment of markers. - HMM model as implemented in 'polyqtlR' - Heuristic model as implmented in 'polyqtlR'

Usage

calc_IBD(
  geno,
  p1hom,
  p2hom,
  map = NULL,
  ploidy = 2,
  method = "naive",
  p1name = "P1",
  p2name = "P2",
  hmm.error = 0.01
)

Arguments

geno	integer or matrix containing genotypes (genotypes of allele A, from 0 to ploidy). If 'geno' is a matrix, each row must be labelled with a marker name, and each column is interpreted as an individual. Note that for many markers of one single individual, a 1 column matrix must be provided.
p1hom	numeric vector or matrix indicating which parental homologue, from parent 1, contains allele A. If 'geno' is an integer, 'p1hom' must be a vector of length ploidy. If 'geno' is a matrix, 'p1hom' must be a matrix of ncol = ploidy and nrow = nrow(geno) and the same rownames as 'geno'.
p2hom	numeric vector or matrix indicating which parental homologue, from parent 2, contains allele A. If 'geno' is an integer, 'p2hom' must be a vector of length ploidy. If 'geno' is a matrix,'p2hom' must be a matrix of ncol = ploidy and nrow = nrow(geno) and the same rownames as 'geno'.
map	data.frame containing "position" and "marker" columns. Required for "hmm" and "heuristic" methods
ploidy	numeric, indicating the ploidy of an organism. Only even ploidies allowed and all individuals are expected to have the same ploidy.
method	character indicating IBD method to use. Either "naive" (default), "hmm" for hidden markov model or "heuristic" for the heuristic model of 'polyqtlR'
p1name	character, column name containing the first parent name in the geno matrix. Required for "heuristic" and "hmm" methods
p2name	character, column name containing the second parent name in the geno matrix. Required for "heuristic" and "hmm" methods
hmm.error	numeric, if "hmm" method is used, the estimated error percentage in the genotypes. The default (0.01) performs quite well if there's not a large number of errors.

Details

It requires homologue matrices ('p1hom' and 'p2hom') indicating which homologues of each parent carry a genotype of 1 or 0 for the tracked allele. Based on this, a probability matrix is returned: for each individual how likely it is, at each marker, that it has inherited each of the homologues. As a generic method, it can use a single genotype of one marker (and two vectors of homologue assignment); a vector of markers (and two homologue matrices) or a matrix of markers (and two homologue matrices).

Value

The naive probability of having inherited each parental homologue. If the offspring genotype is impossible given the parental homologues, NA is returned. If 'geno'is a vector, returns a one-row matrix, each column containing the probability of having inherited that parental homologue (the first columns corresponding to parent 1, and the rest to parent 2). If 'geno' is a matrix, a list of matrices is returned. Each element is the probabilities for all individuals of having inherited one parental homologue. There should be ploidy*2 elements in the list.

Examples

data("genotype")
geno <- geno[,-1:-2] #we take out parental genotypes
data("homologue")

#For a single genotype
IBD <- naive_IBD(geno[1,1],hom[1,1:2],hom[1,3:4], ploidy = 2)
#For a single individual
#The genotype must be a matrix of one column!
IBD <- naive_IBD(geno[,1,drop = FALSE],hom[,1:2],hom[,3:4], ploidy = 2)
#For all individuals
IBD <- naive_IBD(geno,hom[,1:2],hom[,3:4], ploidy = 2)

Alethere/SmoothDescent documentation built on Oct. 21, 2023, 7:11 a.m.