An explanation for the sister repulsion phenomenon in Patterson’s f-statistics

Abstract

Patterson’s f-statistics are among the most heavily utilized tools for analysing genome-wide allele frequency data for demographic inference. Beyond studying admixture,f3andf4statistics are also used for clustering populations to identify groups with similar histories. However, previous studies have noted an unexpected behaviour of f-statistics: multiple populations from a certain region systematically show higher genetic affinity to a more distant population than to their neighbours, a pattern that is mismatched with alternative measures of genetic similarity as well as with spatial relationships. We call this counter-intuitive pattern “neighbour repulsion”. Here, we first present a novel instance of neighbour repulsion, where genomes from Bronze Age East Anatolian sites show higher affinity towards Bronze Age Greece rather than their neighbours. This is observed both usingf3- andf4-statistics, contrasts with archaeological/historical expectation, and also contradicts genetic affinity patterns captured using PCA or MDS on genetic distances. We then propose a simple demographic model to explain this pattern, where the neighbouring populations (such as East Anatolia) receive independent gene flow from an external source after their split from each other. Calculatingf3- andf4-statistics using simulated genetic data with varying population genetic parameters, we confirm that low-level gene flow from an external source into populations from one region, as well as unidirectional gene flow between the studied regions, can create neighbour repulsion in f-statistics. Meanwhile, similar to our empirical observations, an MDS analysis of genetic distances can still capture common demographic history between neighbours. Overall, our results highlight the intricate effects of unaccounted population dynamics when inferring shared demographic history using f-statistics.

Wide-scale Geographical Analysis of Genetic Ancestry in the South African Coloured Population

“…However, the f4-statistics in the form f4 (Chimp, SAC, YRI AFR, Zulu) indicate a greater genetic influence from the West-African Yoruba (Supplementary Figure 8). This mismatch with the ADMIXTURE results has been observed in other studies as well, and is called “neighbour repulsion” [53], where the neighbouring populations (Zulu in this case) received independent gene flow from an external source after their split from the West-Africans. In Afrikaners, the West African non-Bantu Niger-Congo speaker associated component contributes more than the Bantu-speaker associated component [19].…”

“…However, the f4-statistics in the form f4 (Chimp, SAC, YRI AFR, Zulu) indicate a greater genetic influence from the West-African Yoruba (Supplementary Figure 8). This mismatch with the ADMIXTURE results has been observed in other studies as well, and is called "neighbour repulsion" [53],…”