fst: Calculate FST from a population-level kinship matrix or...

In StoreyLab/popkin: Estimate Kinship and FST under Arbitrary Population Structure

Calculate FST from a population-level kinship matrix or vector of inbreeding coefficients

Description

This function simply returns the weighted mean inbreeding coefficient. If weights are NULL (default), the regular mean is calculated. If a kinship matrix is provided, then the inbreeding coefficients are extracted from its diagonal using inbr() (requires the diagonal to contains self-kinship values as popkin() returns, and not inbreeding coefficients as inbr_diag() returns). If there is local inbreeding and it can be estimated (from known pedigrees, for example), it can be subtracted from the total inbreeding coefficients, resulting in a vector of structural inbreeding that correctly averages into FST.

Usage

fst(x, weights = NULL, x_local = NULL)

Arguments

x	The vector of inbreeding coefficients, or the kinship matrix if x is a matrix.
weights	Weights for individuals (optional, defaults to uniform weights)
x_local	An optional vector of inbreeding coefficients, or a local kinship matrix if x_local is a matrix.

Details

The returned weighted mean inbreeding coefficient equals the generalized FST if all individuals are "locally outbred" (i.e. if the self-relatedness of every individual stems entirely from the population structure rather than due partly to having unusually closely related parents, such as first or second cousins). Note most individuals in population-scale human data are locally outbred. If there are locally-inbred individuals, but their local inbreeding cannot be estimated, then the returned value will overestimate FST. Good estimates of local inbreeding can be passed (parameter x_local), in which case the code will subtract their effect and FST will be more accurate.

Value

FST

Examples

# Get FST from a genotype matrix

# Construct toy data
X <- matrix(c(0,1,2,1,0,1,1,0,2), nrow = 3, byrow = TRUE) # genotype matrix
subpops <- c(1,1,2) # subpopulation assignments for individuals

# NOTE: for BED-formatted input, use BEDMatrix!
# "file" is path to BED file (excluding .bed extension)
## library(BEDMatrix)
## X <- BEDMatrix(file) # load genotype matrix object

# estimate the kinship matrix "kinship" from the genotypes "X"!
kinship <- popkin(X, subpops) # calculate kinship from X and optional subpop labels
weights <- weights_subpops(subpops) # can weigh individuals so subpopulations are balanced
Fst <- fst(kinship, weights) # use kinship matrix and weights to calculate fst

Fst <- fst(kinship) # no (or NULL) weights implies uniform weights

inbr <- inbr(kinship) # if you extracted inbr for some other analysis...
Fst <- fst(inbr, weights) # ...use this inbreeding vector as input too!

StoreyLab/popkin documentation built on Jan. 10, 2023, 1:39 p.m.