Likelihood-based test for the consanguinity among samples (InbreedingCoeff)
Category Variant Annotations
OverviewLikelihood-based test for the consanguinuity among samples
This annotation estimates whether there is evidence of consanguinuity in a population. The higher the score, the higher the chance that some samples are related. If samples are known to be related, a pedigree file can be provided so that the calculation is only performed on founders and offspring are excluded.
The output is the inbreeding coefficient 'F' (fixation) statistic, which for large sample sizes converges to the probability that an individual's two alleles are identical by descent, provided that consanguinity is the only source of deviation from Hardy-Weinberg equilibrium. If this assumption is not true F may be negative and the excess heterozygosity often indicates an artifactual variant. It is calculated as F = 1 - (# of het genotypes)/(# of het genotypes expected under Hardy-Weinberg equilibrium). The number of het genotypes expected under Hardy-Weinberg equilibrium is 2*(# of samples)*(ref allele frequency)*(alt allele frequency), where allele frequencies are calculated from the samples' genotypes.
- The Inbreeding Coefficient annotation can only be calculated for cohorts containing at least 10 founder samples.
- The Inbreeding Coefficient annotation can only be calculated for diploid samples.
- The implementation here uses likelihoods rather than counts, so computed InbreedingCoeff values can be sensitive to small changes in the low GQ range, as might happen after "reblocking" low coverage sites, as well as GQ0 versus no-call changes
ExcessHet also describes the heterozygosity of the called samples, giving a probability of excess heterozygosity being observed
GATK version 126.96.36.199-SNAPSHOT built at Thu, 16 Dec 2021 11:57:48 -0800.