P1-11: Detecting and understanding selection in farm animals - a coalescent approach

PhD student: Rebekka Brink-Spalink
Thesis Committee: Prof. Anja Sturm, Prof. Henner Simianer, Prof. Tatyana Krivobokova
Graduation Date: 1/2015

Thesis: online publication

Classical coalescent processes like the Kingman coalescent make some strong assumptions which are often violated in animal breeding. For example, convergence of the Wright-Fisher model to the Kingman coalescent relies on a small probability for an sample in large population to have a common ancestor in the previous generation, i.e. small offspring variability. This cannot be assumed if one animal is artificially selected to provide a large fraction of the next generation. Then there is the possibility for multiple mergers in the limit, while the Kingman coalescent yields a binary tree.

Considering multiple mergers also allows for the treatment of a non-constant population size, which may cause so called population bottlenecks: If the parent generation is considerably smaller than the offspring generation, the chance for more than two individuals to have a common parent may no longer be ignored.

Furthermore, recombination and is not considered in simple models. Recombination leeds to a limiting graph which is not a tree since every individual may have more than one ancestor in the preceeding generation. This graph is the ancestral recombination graph. Also mutation may be included in the model.

To deal with these problems the investigation of more general population models - like the Lambda coalescent and multiple merger models - is necessary.

back to overview of all projects