skater (version 0.1.2)

ibd2kin: Compute kinship coefficient from IBD segments

Description

This function is used to retrieve a relatedness measure from IBD segments. The relatedness value returned is the kinship coefficient.

Usage

ibd2kin(.ibd_data, .map, type = NULL)

Value

Tibble with three columns:

1. id1 (sample identifier 1)

2. id2 (sample identifier 2)

3. kinship (kinship coefficent derived from shared segments)

Arguments

.ibd_data

Tibble with IBD segments created using the read_ibd function

.map

Tibble with the genetic map data created using the read_map function

type

Type of IBD to use for kinship coefficient calculation; must be 'IBD1', 'IBD2', or NULL (both IBD1 and IBD2 will be treated the same); default is NULL

Details

The input data should be pairwise IBD segments prepared via read_ibd. The function will internally loop over each chromosome, and use a specified genetic map to convert shared segments to genetic units. After doing so, the function converts the shared length to a kinship coefficient by summing \(0.5*IBD2 + 0.25*IBD1\).

Note that the data read in by read_ibd when source="pedsim" returns a list with separate tibbles for IBD1 and IBD2 segments. The current implementation of this function requires running this function independently on IBD1 and IBD2 segments, then summarizing (adding) the corresponding proportions. See examples.

References

http://faculty.washington.edu/sguy/ibd_relatedness.html

Examples

pedsim_fp <- system.file("extdata", "GBR.sim.seg.gz", package="skater", mustWork=TRUE)
pedsim_seg <- read_ibd(pedsim_fp, source = "pedsim")
gmapfile <- system.file("extdata", "sexspec-avg-min.plink.map", package="skater", mustWork=TRUE)
gmap <- read_map(gmapfile)
ibd1_dat <- ibd2kin(.ibd_data=pedsim_seg$IBD1, .map=gmap, type="IBD1")
ibd2_dat <- ibd2kin(.ibd_data=pedsim_seg$IBD2, .map=gmap, type="IBD2")
dplyr::bind_rows(ibd1_dat,ibd2_dat) %>%
  dplyr::group_by(id1,id2) %>%
  dplyr::summarise(kinship = sum(kinship), .groups = "drop")