Rannsóknastofa í þróunarfræði og stofnerfðafræði við Líffræðistofnun Háskóla Íslands boðar til málþings
í stofu 204 í Lögbergi þann 20. september kl 13:15–15:30
Seminar Lögberg room 204 September 20, 13:15–15:30
Sweepstakes reproduction and reticulations: genetics and genomics of some cod-fish
Einar Árnason and Katrín Halldórsdóttir, University of Iceland
In his Sex and Evolution book Williams discusses the „Cod-Starfish model“ of sweepstakes reproductive success. Cod are among the most fecund vertebrates. We are using genetics and genomics to study cod-fish reproduction and present evidence for reproductive sweepstakes in parameter estimates using models of multiple merger coalescence. Cod are major players in the ecosystem and have an interesting trans-Arctic biogeography from the Atlantic into the Pacific to which they also have
adapted. We ask about codweb, the nature of adaptive variation using genomic phylonetwork analysis.
A structured model for population growth or multiple merger coalescents
Maite Wilke Berenguer, Technische Universität Berlin
We introduce a model for a diploid, bi-parental structured population evolving in discrete, nonoverlaping generations which (potentially) includes the phenomena of population growth or multiple mergers as well as mutation, selection and recombination and identify the conditions on the parameters involved under which the model converges to a structured -ancestral influence graph or a structured Kingman coalescent with population growth.
Beta-Xi-Sim and the singleton tail statistic for model selection of multiple mergers
Jere Koskela, Technische Universität Berlin
I will introduce Beta-Xi-Sim as an open source simulator of diploid, biparental Xi-coalescent models under complex evolutionary scenarios involving the joint effect of crossover recombination, weak natural selection and discrete spatial structure with migration. It comes with a user-friendly interface, and can flexibly model scenarios such as arbitrary numbers of unlinked loci, arbitrary recombination maps at each locus, and geography-dependent diploid selection. I will then introduce the singleton-tail statistic: a subset of the site frequency spectrum, and present simulation results which demonstrate that it discriminates between Xi-coalescents and population growth models based on the Kingman coalescent with very high power. Finally, I will demonstrate that this model selection can be done in the presence of unknown recombination or weak selection with no loss of power, but that correctly specifying the spatial structure of the population is vital to the performance of the method.
Modeling gene genealogies in highly fecund populations
Bjarki Eldon, Museum für Naturkunde, Berlin.
Highly fecund populations abound in nature, yet traditional mathematical population genetic
models do not capture high fecundity. Only very recently have models begun to emerge which
explicitly formulate high fecundity. Of immediate concern is how to detect high fecundity in genetic
and genomic data. We address this problem from two angles; by considering the gene genealogies
of nested subsamples, and by considering the impact of large sample size on our coalescent model.
Our results identify limits of population genetic data in distinguishing between models. This is
joint work with Fabian Freund and Alison Etheridge.
Einar Árnason (einararn at hi.is), prófessor í þróunar- og stofnerfðafræði og Katrín Halldórsdóttir (katrinh at hi.is) sérfræðingur.