Academic journal article Genetics

Inferring Recent Demography from Isolation by Distance of Long Shared Sequence Blocks

Academic journal article Genetics

Inferring Recent Demography from Isolation by Distance of Long Shared Sequence Blocks

Article excerpt

(ProQuest: ... denotes formulae omitted.)

THERE has been a long-standing interest in estimating demography, as migration and population density are key parameters for studying evolution and ecology. Demographic models are essential for disentangling the effects of neutral evolution from selection, and are crucial to understanding local adaptation. Moreover, the inference of demographic parameters is important for conservation and breeding management. Given the intensive nature of obtaining such parameters by direct observations, which are moreover necessarily limited to short timescales, the increasing availability of genetic markers has spurred efforts to develop inference methods based on genetic data.

This work focuses on estimating dispersal rate and population density in two-dimensional habitats by analyzing the geographic distribution of so-called identity-by-descent (IBD) blocks, which are commonly defined as co-inherited segments delimited by recombination events (see Figure 1). It has now become feasible to detect long regions of exceptional pairwise similarities from dense SNP or whole genome sequences (Gusev et al. 2009; Browning and Browning 2011). For regions longer than a few cM, the bulk mostly consists of a single IBD block unbroken by recombination, at least when inbreeding is rare (Chiang et al. 2016). This yields novel opportunities for inferring recent demography, as one can study the direct traces of coancestry.

Moreover, the length of shared blocks contains information about their age. That is, the longer the time to the most recent common ancestor, the shorter the expected IBD length, as recombination has more chances to break up ancestral genetic material. The probability that no recombination occurs in a block of a given map length decays exponentially, going back in time. Hence, long IBD blocks originate mostly from very recent coancestry and provide insight into the recent history of a population. Shared long blocks between pairs of populations can be used to infer the distribution of recent coalescence times (Ralph and Coop 2013), and fitting deme and island models can yield information on recent population sizes (Palamara et al. 2012; Browning and Browning 2015) and migration patterns (Palamara and Pe'er 2013). These works are complementary to the analysis of short identical segments, which are informative about deeper timescales (Li and Durbin 2011; Harris and Nielsen 2013), and they showcase the utility of long IBD blocks for inferring recent demography.

Here, we focus on a pattern of isolation by distance of IBD blocks within populations extended in two dimensions with local migration. For such populations, the classical WrightMalécot formula describes an increase of mean pairwise genetic diversity with increasing geographic separation (Wright 1943; Malécot 1948). Several inference methods use such classical isolation by distance patterns as signals to infer the parameters of recent demography. For example, fitting increasing pairwise genetic diversity with geographic distance is widely used (Rousset 1997, 2000; Vekemans and Hardy 2004), and approximate-Bayesiancomputation methods have been applied (Joseph et al. 2016). Similarly, the extent of geographic clustering of rare frequency alleles can be used as a signal for inference (Novembre and Slatkin 2009). While the signal of locally decreased pairwise genetic diversity mostly stems from recent times (Leblois et al. 2004), such patterns can be severely confounded by deeper, often unknown, ancestral patterns (Meirmans 2012). Moreover, such methods can usually only infer the neighborhood size 4pDes2, which is proportional to the product of dispersal rate s2 with effective density De. Usually, these important parameters cannot be estimated separately, as the underlying signal is mostly based on a short-term equilibrium between local drift and dispersal. An exception is quickly mutating organisms such as viruses, for which phylogeographic diffusion approaches yield separate estimates of s (Lemey et al. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.