Identity by descent in population data

Abstract
Two individuals’ DNA is said to be identical by descent when the individuals inherited it from a shared
ancestor such as a great-grandparent. Identity by descent (IBD) occurs in relatively long segments of
several million base pairs when the shared ancestor lived within approximately the past 50 generations.
Since ancestors from 50 generations ago can have many descendants, it is possible to find segments of
IBD between many pairs of individuals without known relationships.
The goal of this proposal is to develop three new applications of IBD segments.
The first application is estimating recombination rates. Recombination rates vary across the genome, and
the rates must be estimated in order to perform many other genetic analyses. We will develop methods
that use IBD segments to estimate recombination rates from population sample data. We will also develop
methods to rigorously compare IBD-based recombination maps across populations. We will estimate
recombination rates in different human populations, and we will investigate regions of the genome where
there are differences between populations.
The second application is estimating sex-specific demographic history. Males and females can have
different effective population sizes due to differing variability in reproductive success, and males and
females often have different rates of migration. We will develop method to estimate sex-specific effective
population sizes and mutation rates by comparing rates of IBD segments on the X chromosome (which is
weighted towards female history) and the autosomes (which are equally weighted between male and
female histories). This work will improve our understanding of the historical dynamics that have shaped
populations.
The third application is analyzing recent positive selection. Selected regions can be found by looking for
elevated levels of IBD. We will develop statistical methodology to estimate the strength of selection, the
mode of selection, and the frequency of the selected variants. We know that selection differs across
geographic regions due to different pathogens and other environmental exposures, and we will analyze
diverse populations to catalog selected regions and how they differ between populations.

Grant Number: 5R01HG005701-11
NIH Institute/Center: NIH
Principal Investigator: BRIAN BROWNING