nadiv (version 2.18.0)

ggcontrib: Genetic group contribution

Description

Calculates the genomic contribution each genetic group makes to every individual in a pedigree

Usage

ggcontrib(pedigree, ggroups = NULL, fuzz = NULL, output = "matrix")

Value

Returns i x r genetic group contributions to all 'i' individuals from each of the 'r' genetic groups. Default output is an object of class

matrix (dense), but this format can be changed (e.g., "dgCMatrix" for a sparse matrix).

Arguments

pedigree

A pedigree where the columns are ordered ID, Dam, Sire

ggroups

An optional vector of either: genetic group assignment for every individual or just the unique genetic groups. fuzz must be NULL if an object is supplied to the ggroups argument.

fuzz

A matrix containing the fuzzy classification of phantom parents into genetic groups. ggroups must be NULL if an object is supplied to the fuzz argument.

output

Format for the output

Author

Details

The specification of genetic groups is done in one of two approaches, either using fuzzy classification or not.

Fuzzy classification enables phantom parents to be assigned to (potentially) more than one genetic group (Fikse 2009). This method requires unique phantom parent identities to be included in the pedigree for all observed individuals with unknown parents. For 'p' phantom parents, 'p' identities should be listed as individuals in the first 'p' rows of the pedigree and these should be the only individuals in the pedigree with missing values in their Dam and Sire columns (denoted by either 'NA', '0', or '*'). The matrix supplied to the fuzz argument should have 'p' rows (one for each phantom parent) and 'r' columns, where 'r' is the number of genetic groups.

Non-fuzzy classification can handle the specification of genetic groups in three formats:

(1) similar to ASReml's format for specifying genetic groups, the first 'r' rows of the pedigree (given to the pedigree argument) contain the label for each genetic group in the ID column and indicate missing values for the Dam and Sire columns (denoted by either 'NA', '0', or '*'). No object is supplied to the ggroups argument. All individuals in the pedigree must then have one of the 'r' genetic groups as parent(s) for each unknown parent. Note, a warning message indicating In numPed(pedigree): Dams appearing as Sires is expected, since the dam and sire can be the same for all individuals in the pedigree composing the base population of a genetic group.

(2) similar to Jarrod Hadfield's rbv function arguments in the MCMCglmm package, for a pedigree of dimension i x 3 (given to the pedigree argument), where 'i' is the total number of individuals in the pedigree, a similar vector of length 'i' is given to the ggroups argument. This vector lists either the genetic group to which each individual's phantom parents belong or NA if the individual is not to be considered part of one of the base populations (genetic groups). NOTE, this approach does not allow phantom dams and phantom sires of a given individual to be from different genetic groups.

(3) similar to DMU's format for specifying genetic groups. For a pedigree of dimension i x 3 (given to the pedigree argument), where 'i' is the total number of individuals in the pedigree, instead of missing values for a parent, one of the 'r' genetic groups is specified. A character vector of length 'r' with unique genetic group labels is given to the ggroups argument. Note, that all individuals with a missing parent should have a genetic group substituted instead of the missing value symbol (i.e., either 'NA', '0', or '*').

References

Fikse, F. 2009. Fuzzy classification of phantom parent groups in an animal model. Genetics, Selection, Evolution. 41:42.

Quaas, R.L. 1988. Additive genetic model with groups and relationships. Journal of Dairy Science. 71:1338-1345.

Examples

# Use the pedigree from Quaas 1988 (See `data(Q1988)`)
##########################
# Fuzzy classification
  ## Fuzzy classification with complete assignment to one group
    Q1988fuzz <- Q1988[-c(1:2), c("id", "phantomDam", "phantomSire")]
    Qfnull <- matrix(c(1,0,0,1,0, 0,1,1,0,1), nrow = 5, ncol = 2,
	dimnames = list(letters[1:5], c("g1", "g2")))
    (Qfuzznull <- ggcontrib(Q1988fuzz, fuzz = Qfnull))

    ## Should be identical to the non-fuzzy classification output
    # format (1) from above
      (Q <- ggcontrib(Q1988[-c(3:7), c(1,4,5)]))
    stopifnot(Qfuzznull == Q)

  ## Fuzzy classification with arbitrary assignments
    Qf <- matrix(c(1,0,0.5,0.5,0.5, 0,1,0.5,0.5,0.5), nrow = 5, ncol = 2,
	dimnames = list(letters[1:5], c("g1", "g2")))
    (Qfuzz <- ggcontrib(Q1988fuzz, fuzz = Qf))  

  ## Using the pedigree and fuzzy classification in Fikse (2009)
    F2009fuzz <- data.frame(id = c(letters[1:7], LETTERS[1:6]),
	dam = c(rep(NA, 7), "a", "c", "e", "A", "C", "D"),
	sire = c(rep(NA, 7), "b", "d", "f", "B", "g", "E"))
    Ff <- matrix(c(1,0,1,0,0,0,0.2,
		0,1,0,0.6,0,0.3,0.4,
		0,0,0,0.4,1,0.7,0.4),
		nrow = 7, ncol = 3,
		dimnames = list(letters[1:7], paste0("g", 1:3)))
    # Actual Q matrix printed in Fikse (2009)
      Fikse2009Q <- matrix(c(0.5,0.5,0,0.5,0.1,0.3,
			0.5,0.3,0.15,0.4,0.275,0.3375, 
			0,0.2,0.85,0.1,0.625,0.3625),
		nrow = 6, ncol = 3,
		dimnames = list(LETTERS[1:6], paste0("g", seq(3))))

    Ffuzz <- ggcontrib(F2009fuzz, fuzz = Ff)
      (diffFfuzz <- Ffuzz - Fikse2009Q)
      # Encountering some rounding error
      stopifnot(length((drop0(diffFfuzz, tol = 1e-12))@x) == 0)


##########################
# Non-fuzzy classification
  # format (1) from above
    Q1 <- Q1988[-c(3:7), c(1,4,5)]
    (gg1 <- ggcontrib(Q1, ggroups = NULL)) # note the warning message which is typical

  # format (2) from above
    Q2 <- Q1988[-c(1:7), 1:3]
    # arbitrarily assign individuals genetic groups for unknown parents
    ## Means gg2 is NOT comparable to gg1 or gg3!
    ggvec.in <- c("g1", "g2", "g1", NA)
    (gg2 <- ggcontrib(Q2, ggroups = ggvec.in))

  # format (3) from above
    Q3 <- Q1988[-c(1:7), c(1,4,5)]
    gg3 <- ggcontrib(Q3, ggroups = c("g1", "g2"))

  stopifnot(gg1 == gg3)