Measuring individual inbreeding in the age of genomics: marker-based measures are better than pedigrees

Abstract

Inbreeding (mating between relatives) can dramatically reduce the fitness of offspring by causing parts of the genome to be identical by descent. Thus, measuring individual inbreeding is crucial for ecology, evolution and conservation biology. We used computer simulations to test whether the realized proportion of the genome that is identical by descent (IBD G ) is predicted better by the pedigree inbreeding coefficient (F P ) or by genomic (marker-based) measures of inbreeding. Genomic estimators of IBD G included the increase in individual homozygosity relative to mean Hardy-Weinberg expected homozygosity (F H ), and two measures (F ROH and F E ) that use mapped genetic markers to estimate IBD G . IBD G was more strongly correlated with F H , F E and F ROH than with F P across a broad range of simulated scenarios when thousands of SNPs were used. For example, IBD G was more strongly correlated with F ROH , F H and F E (estimated with ⩾ 10 000 SNPs) than with F P (estimated with 20 generations of complete pedigree) in populations with a recent reduction in the effective populations size (from N e = 500 to N e = 75). F ROH , F H and F E generally explained 490% of the variance in IBD G (among individuals) when 35 K or more SNPs were used. F P explained o80% of the variation in IBD G on average in all simulated scenarios, even when pedigrees included 20 generations. Our results demonstrate that IBD G can be more precisely estimated with large numbers of genetic markers than with pedigrees. We encourage researchers to adopt genomic marker-based measures of IBD G as thousands of loci can now be genotyped in any species.

Recent advances in conservation and population genomics data analysis

“…Genomic approaches capture variation in realized inbreeding that is missed by pedigree analysis due to the stochastic effects of linkage and unknown common ancestors of parents (Franklin, 1977; Thompson, 2013). Thus, while deep and accurate pedigrees can often precisely measure individual inbreeding in species with many chromosomes and/or high recombination rates (Kardos et al., 2018; Knief, Kempenaers, & Forstmeier, 2017; Nietlisbach et al., 2017), genomic approaches are expected to more reliably measure inbreeding and inbreeding depression (Kardos, Luikart, & Allendorf, 2015a; Kardos et al., 2018; Keller, Visscher, & Goddard, 2011; Wang, 2016). Given that many studies have used only shallow pedigrees or few DNA markers, it is possible that power to detect inbreeding depression has been low; therefore, inbreeding depression could be more common, widespread, and severe than previously thought.…”

Meiotic recombination shapes precision of pedigree- and marker-based estimates of inbreeding

“…Thus, the longer the genetic map of a genome, which directly corresponds with the expected number of crossovers in each meiosis, the smaller the Mendelian sampling noise (Hill and Weir, 2011;Wang, 2016). To our knowledge, all analytical analyses and most of the simulations so far have assumed a uniform distribution of crossovers along chromosomes (see, for example, Franklin, 1977;Stam, 1980;Hill and Weir, 2011;Kardos et al, 2015;Wang, 2016; but see Suarez et al, 1979 andLibiger andSchork, 2007 for Monte Carlo simulations on relatedness). Although this assumption holds more or less for the human genome (Matise et al, 2007), linkage maps from other species have shown that the distribution of recombination along chromosomes can be highly biased toward the telomeres (Gore et al, 2009;Backström et al, 2010).…”

“…This will introduce variation into our estimates of GWIBD which we call the marker sampling noise. Unlike the Mendelian sampling noise that is an inherent result of the meiotic process, the marker sampling noise is only a consequence of the limited number of markers used; hence, the precision of GWIBD estimates increases when more molecular markers are sampled Kardos et al, 2015;Wang, 2016). Meiotic recombination also influences Marker F (Wang, 2016): given a fixed number of markers, the precision of Marker F decreases with the amount of meiotic recombination (whereas precision of Pedigree F increases), because each marker contains less information about GWIBD in genomes with more independently segregating segments (that is, longer genetic maps; Wang, 2016).…”

“…On the other hand, Marker F reflects the actual proportion of GWIBD of an individual. A multitude of methods exist for calculating Marker F but until now no consensus has been reached on which of these methods yields the most precise estimate of GWIBD (a first evaluation of the different metrics was performed in Kardos et al, 2015). Given a large number of genotyped molecular markers (that is, ⩾ 10 000 singlenucleotide polymorphisms (SNPs)), both averages of single marker estimates and estimates based on several closely linked markers perform equally well in predicting GWIBD (Kardos et al, 2015).…”

“…Depending on how well marker homozygosity reflects IBD, this may introduce error into the GWIBD estimate (Thompson, 1976; that we call the IBD-IBS discrepancy. We here specifically refer to a limited number of microsatellites for GWIBD estimation because both the marker sampling noise and the IBD-IBS discrepancy cease when large numbers (⩾10 000) of SNPs are employed for GWIBD estimation Kardos et al, 2015;Wang, 2016).…”