High level of inbreeding in final phase of 1000 Genomes Project

Abstract

The 1000 Genomes Project provides a unique source of whole genome sequencing data for studies of human population genetics and human diseases. The last release of this project includes more than 2,500 sequenced individuals from 26 populations. Although relationships among individuals have been investigated in some of the populations, inbreeding has never been studied. In this article, we estimated the genomic inbreeding coefficient of each individual and found an unexpected high level of inbreeding in 1000 Genomes data: nearly a quarter of the individuals were inbred and around 4% of them had inbreeding coefficients similar or greater than the ones expected for first-cousin offspring. Inbred individuals were found in each of the 26 populations, with some populations showing proportions of inbred individuals above 50%. We also detected 227 previously unreported pairs of close relatives (up to and including first-cousins). Thus, we propose subsets of unrelated and outbred individuals, for use by the scientific community. In addition, because admixed populations are present in the 1000 Genomes Project, we performed simulations to study the robustness of inbreeding coefficient estimates in the presence of admixture. We found that our multi-point approach (FSuite) was quite robust to admixture, unlike single-point methods (PLINK).

Figure 1. Accuracy of inbreeding estimators in simulated admixed samples.

The differences between estimated and true f values (Δf) and the genomic proportions of European ancestry (ADM_CEU) were calculated on one random individual (1C, 2C and OUT) from 100 sample replicates (total 100 per mating type). Only FSuite estimates with Q >50 were plotted and single-point negative estimates (PLINK and REAP) were set to 0. Four sets of allele frequencies were used for FSuite and PLINK: European (CEU), African (YRI) and Asian (JPT/CHB) reference frequencies, and frequencies estimated on each sample (SAMPLE). REAP used individual allele frequencies. 1C = first-cousin offspring; 2C = second-cousin offspring; OUT = outbred individual.

Figure 2. Inbreeding estimation and detection in TGP populations.

Each point represents the f estimate for one individual. Large points represent the ones that are inferred as offspring of first-cousin (1C) or closest relationships, medium open points, the ones that are inferred as offspring of second-cousin offspring (2C), and small points, the ones that are inferred as outbred. Individuals were ordered in each population according to their f values. See Table 1 for the description of the different populations.