A general approach for inferring the ancestry of recent ancestors of an admixed individual

Abstract

The genome of an individual from an admixed population consists of segments originated from different ancestral populations. Most existing ancestry inference approaches focus on calling these segments for the extant individual. In this paper, we present a general ancestry inference approach for inferring recent ancestors from an extant genome. Given the genome of an individual from a recently admixed population, our method can estimate the proportions of the genomes of the recent ancestors of this individual that originated from some ancestral populations. The key step of our method is the inference of ancestors (called founders) right after the formation of an admixed population. The inferred founders can then be used to infer the ancestry of recent ancestors of an extant individual. Our method is implemented in a computer program called PedMix2. To the best of our knowledge, there is no existing method that can practically infer ancestors beyond grandparents from an extant individual's genome. Results on both simulated and real data show that PedMix2 performs well in ancestry inference.

Keywords: ancestry inference; genetic tests; genetics; population admixture; recombination.

Fig. 1.

Example of a perfect pedigree with $K=3$ ancestral populations with four generations ($g=4$). The founders are on the top. At each generation (viewing upward), the number of ancestors doubles. The red lines show an inheritance path for a site. A single extant haplotype $H_{\mathit{pq}}$ is at the bottom. Each founder is labeled by one of three ancestral populations (A, B, and C) that form the admixed population. The list of these labels forms the founder configuration.

Fig. 2.

Empirical investigation of the relationship between founder ratio and admixture proportion. Part 2 (A): absolute mean differences between the true admixture proportions and the true founder ratios with varying numbers of chromosomes (under 95% CI). The differences converge to $1\%$ when there were $100$ chromosomes, which provides empirical justification of Proposition 1. Part 2 (B): comparison of PedMix2 with RFMix in the accuracy of admixture proportion and founder ratio accuracy of PedMix2 with different numbers of chromosomes. The default settings in Table 1 are used. “red +”: the mean value.

Fig. 3.

Comparison of PedMix2 with RFMix and PedMix on admixture proportion accuracy for the extant individual and all $10$ generations of ancestors. Data simulated under the default settings in Table 1. $g=11$, “red +”: the mean value.

Fig. 4.

Accuracy of inferred admixture proportions (of extant individuals and the ancestors within three generations) under various $\rho ,t,\mu$, and $n_b$ values. (A) Different recombination rates $\rho$. (B) Different splitting time t. (C): different mutation rates $\mu$. (D): different numbers of blocks $n_b$. $t$: six values ($0.001$, $0.01$, $0.15$, $0.3$, $0.5$, and $0.9$ coalescent unit). “g-grandparent”: great grandparent. “red +”: the mean value.

Fig. 5.

Running time (in natural log) of PedMix2 under varying the number of generations in the pedigree and the number of blocks per chromosome. For the experiments with three reference populations, the number of blocks per chromosome ($n_b$) is set to $20$.

Fig. 6.

Admixture proportion accuracy of extant individuals and their parents/grandparents for $10$ trios from the ASW population of the 1000 Genomes Project. (A) Accuracy of PedMix2 for extant individual inference. (B) (parents) and (C) (grandparents): compare PedMix2 and PedMix with data that are trimmed for estimating parental and grandparental admixture proportion accuracy. “red +”: the mean value.