qtl2 (version 0.24)

calc_kinship: Calculate kinship matrix

Description

Calculate genetic similarity among individuals (kinship matrix) from conditional genotype probabilities.

Usage

calc_kinship(
  probs,
  type = c("overall", "loco", "chr"),
  omit_x = FALSE,
  use_allele_probs = TRUE,
  quiet = TRUE,
  cores = 1
)

Arguments

probs

Genotype probabilities, as calculated from calc_genoprob().

type

Indicates whether to calculate the overall kinship ("overall", using all chromosomes), the kinship matrix leaving out one chromosome at a time ("loco"), or the kinship matrix for each chromosome ("chr").

omit_x

If TRUE, only use the autosomes; ignored when type="chr".

use_allele_probs

If TRUE, assess similarity with allele probabilities (that is, first run genoprob_to_alleleprob()); otherwise use the genotype probabilities.

quiet

IF FALSE, print progress messages.

cores

Number of CPU cores to use, for parallel calculations. (If 0, use parallel::detectCores().) Alternatively, this can be links to a set of cluster sockets, as produced by parallel::makeCluster().

Value

If type="overall" (the default), a matrix of proportion of matching alleles. Otherwise a list with one matrix per chromosome.

Details

If use_allele_probs=TRUE (the default), we first convert the genotype probabilities are converted to allele probabilities (using genoprob_to_alleleprob()). This is recommended, as then the result is twice the empirical kinship coefficient (e.g., the expected value for an intercross is 1/2; using genotype probabilities, the expected value is 3/8).

We then calculate \(\sum_{kl}(p_{ikl} p_{jkl})\) where \(k\) = position, \(l\) = allele, and \(i,j\) are two individuals.

For crosses with just two possible genotypes (e.g., backcross), we don't convert to allele probabilities but just use the original genotype probabilities.

Examples

# NOT RUN {
grav2 <- read_cross2(system.file("extdata", "grav2.zip", package="qtl2"))
map <- insert_pseudomarkers(grav2$gmap, step=1)
probs <- calc_genoprob(grav2, map, error_prob=0.002)
K <- calc_kinship(probs)

# using only markers/pseudomarkers on the grid
grid <- calc_grid(grav2$gmap, step=1)
probs_sub <- probs_to_grid(probs, grid)
K_grid <- calc_kinship(probs_sub)
# }