Estimating Human Inbreeding Coefficients: Comparison of Genealogical and Marker Heterozygosity Approaches

Abstract

SummaryWe have used genealogies and genomic polymorphisms to estimate individual inbreeding coefficients (F) in 50 subjects with an expected range (based on recent genealogies) of F from 0.0 to 0.0625. The estimates were based on two approaches, using genotypes respectively from 410 microsatellite markers (410-STR panel) and from 10,000 SNPs (10K-SNP panel). The latter was performed in a sub-sample of 15 individuals. We concluded that for both marker panels measures of inbreeding based on the excess of homozygosity over Hardy-Weinberg expectation were not closely correlated with 4-5 generation genealogical F-values. For the 10K-SNP panel we found two alternative measures which correlated more closely with F, based respectively on standard errors and on paired homozygosity of nearby SNPs over distances of 2-4 cM. We propose an empirical method for estimating standard errors and hence individual F-values, based on the variation between individual autosomes. This method could provide useful estimates of average F-values for groups of individuals in population-based studies of the effects of inbreeding/homozygosity on quantitative traits.

American College of Medical Genetics and Genomics: standards and guidelines for documenting suspected consanguinity as an incidental finding of genomic testing

“…In theory, in the offspring of a second-cousin mating, an average of four 12.5 Mb stretches of homozygosity per genome will be present, although both the number and the size of homozygous segments are known to be highly variable. 5 When long contiguous stretches of homozygosity involving multiple chromosomes are present, the percentage of the genome that is IBD can be estimated by the sum of the sizes of the homozygous segments divided by the total autosomal genomic length (~2,881 Mb for GRCh37/hg19). The sex chromosomes are typically excluded from the calculation because males have only a single X and Y chromosome and therefore cannot have homozygosity at any locus outside of the pseudoautosomal regions.…”

“…These theoretical values are found in many genetics texts and resources. 5 Genomic testing, including single-nucleotide polymorphism-based microarrays and whole-genome sequencing, can detect long stretches of the genome that display homozygosity. The presence of these segments, when distributed across multiple chromosomes, can indicate a familial relationship between the proband's parents.…”

“…Even among the progeny of first cousins, in whom the average percentage of the genome that is IBD is 6.25%, the SD is 2.43%. 5 The expected percentages are based on a single common ancestor; however, multiple loops of consanguinity or multiple generations of breeding within a relatively closed community could complicate the estimation of the degree of relationship. These variations from the expected or theoretical values are more pronounced for more distantly related individuals 5 and may be caused by stochastic events, multiple loops of consanguinity, small gene pools, and unknown family structures (e.g., adoptions, nonpaternity).…”

“…5 The expected percentages are based on a single common ancestor; however, multiple loops of consanguinity or multiple generations of breeding within a relatively closed community could complicate the estimation of the degree of relationship. These variations from the expected or theoretical values are more pronounced for more distantly related individuals 5 and may be caused by stochastic events, multiple loops of consanguinity, small gene pools, and unknown family structures (e.g., adoptions, nonpaternity). Because of these variables, the specific familial relationship or degree of relatedness between the parents cannot always be extrapolated from the percentage of the genome that is IBD.…”

Quantifying the increase in average human heterozygosity due to urbanisation

“…15 Second, we used a small subset of individuals from this sample (with varying degrees of consanguinity determined from genealogical records) to define statistical methods that could be used to determine each individual's h-values using both STR and SNP-based genome-wide scans with reasonable precision. 16 Third, we applied this approach to investigate the effects of h on a range of biomedically relevant human quantitative traits with two main conclusions: (i) our estimated h-values showed a very high correlation with genealogical expectations; and (ii) we showed statistically significant associations between h and blood pressure and cholesterol levels. 17 To further investigate our general hypothesis that global outbreeding could have substantial effects on burden of disease, 11 we need to quantify the typical increase in average h that occurs during isolate break-up and urbanisation, which we attempted to do in this study.…”