bnpsd (version 1.0.4)

qis: Construct admixture proportion matrix for independent subpopulations

Description

This function constructs an admixture proportion matrix where every individuals is actually unadmixed (draws its full ancestry from a single intermediate subpopulation). The inputs are the vector of subpopulation labels labs for every individual (length \(n\)), and the length-\(k\) vector of unique subpopulations subpops in the desired order. If subpops is missing, the sorted unique subpopulations observed in labs is used. This function returns the admixture proportion matrix \(Q\) with individuals along the rows and subpopulations along the columns, marking for each individual TRUE for the column corresponding to its subpopulation, FALSE otherwise. Treating the entries of \(Q\) as numerical, every individual has an admixture proportion of 1 for its subpopulation and 0 for all other subpopulations.

Usage

qis(labs, subpops)

Arguments

labs

Length-\(n\) vector of subpopulation labels

subpops

Optional length-\(k\) vector of unique subpopulations in desired order. Stops if subpops does not contain all unique labels in labs (no error if subpops contains additional labels).

Value

The \(n \times k\) admixture proportion matrix \(Q\). The unique subpopulation labels are given in colnames(Q).

Examples

# NOT RUN {
# vector of subpopulation memberships
labs <- c(1,1,1,2,2,3,1)
# admixture matrix with subpopulations (along columns) sorted
Q <- qis(labs)

# declare subpopulations in custom order
subpops <- c(3,1,2)
# columns will be reordered to match subpops as provided
Q <- qis(labs, subpops)

# declare subpopulations with unobserved labels
subpops <- 1:5
# note columns 4 and 5 will be false for all individuals
Q <- qis(labs, subpops)

# }