A universal mechanism generating clusters of differentiated loci during divergence-with-migration. (1st July 2016)

• Record Type: Journal Article
• Authors: Rafajlović, Marina; Emanuelsson, Anna; Johannesson, Kerstin; Butlin, Roger K.; Mehlig, Bernhard
• Is Part Of: Evolution. Volume 70:Number 7(2016)
• Page Start: 1609
• Page End: 1621
• Publication Date: 2016-07-01
• Subjects: Concentrated genetic architecture -- divergence hitchhiking -- divergent selection -- genomic hitchhiking -- islands of divergence -- stochastic loss; Evolution -- Periodicals; Heredity -- Periodicals; Évolution (Biologie) -- Périodiques; Hérédité -- Périodiques; 338.47004094
• DOI: 10.1111/evo.12957
• Languages: English
• ISSNs: 0014-3820

Abstract

Abstract: Genome-wide patterns of genetic divergence reveal mechanisms of adaptation under gene flow. Empirical data show that divergence is mostly concentrated in narrow genomic regions. This pattern may arise because differentiated loci protect nearby mutations from gene flow, but recent theory suggests this mechanism is insufficient to explain the emergence of concentrated differentiation during biologically realistic timescales. Critically, earlier theory neglects an inevitable consequence of genetic drift: stochastic loss of local genomic divergence. Here, we demonstrate that the rate of stochastic loss of weak local differentiation increases with recombination distance to a strongly diverged locus and, above a critical recombination distance, local loss is faster than local "gain" of new differentiation. Under high migration and weak selection, this critical recombination distance is much smaller than the total recombination distance of the genomic region under selection. Consequently, divergence between populations increases by net gain of new differentiation within the critical recombination distance, resulting in tightly linked clusters of divergence. The mechanism responsible is the balance between stochastic loss and gain of weak local differentiation, a mechanism acting universally throughout the genome. Our results will help to explain empirical observations and lead to novel predictions regarding changes in genomic architectures during adaptive divergence.