Thesis
Cultural transmission of fitness in mathematical population genetics
- Abstract:
-
Cultural transmission of fitness is the observed phenomenon in human populations that the number of a parent's siblings in uences the number of their own offspring. We propose an extension of the classical Wright-Fisher model for constant size haploid populations to allow for setting the magnitude of these `in uences' for each family size separately. This is achieved by assigning a reproductive weight to each individual in the parental generation depending on their family size.
The model is a robust extension in the sense that both the forward-in-time model and its dual coalescent process behave similarly to Wright-Fisher and the latter's own limit dual, the n-coalescent, respectively, provided the weight sequence has a constant tail, that is, individuals with many siblings receive the same reproductive weight. Indeed, in the limit of infinite population size, the coalescent dual of our model converges weakly to a constant scaled n-coalescent.
Second, we present results of numerical simulations of the extended Wright- Fisher model with strictly increasing weight sequences of the form w = (jα), α ≥ 0. Here we observe a phase transition in family size frequency at the point α = 1; and we report multiple mergers occurring in the model's coalescent dual. We also use the data to approximate the coalescent's natural time scale. Such a rapidly increasing weight sequence as this one seems to lack biological relevance, however.
Third, we study a near-critical multi-type branching process model of the phenomenon. A sequence of rescaled Galton-Watson branching processes weakly converges to a Feller-type diffusion. The drift and the diffusion coeficients are expressed in terms of the reproductive weight sequence.
Fourth, we describe a branching diffusion model for cultural transmission of fitness whose dual process is the Bolthausen-Sznitman U-coalescent demonstrating that the phenomenon can naturally lead to a whole variety of dual processes with different qualitative features.
Actions
Access Document
- Files:
-
-
(Preview, Dissemination version, pdf, 4.7MB, Terms of use)
-
Authors
Contributors
- Institution:
- University of Oxford
- Division:
- MPLS
- Department:
- Statistics
- Role:
- Supervisor
- Type of award:
- DPhil
- Level of award:
- Doctoral
- Awarding institution:
- University of Oxford
- Language:
-
English
- UUID:
-
uuid:b5848403-1704-49ff-b33f-026a9915bf2e
- Deposit date:
-
2018-03-27
- ARK identifier:
Terms of use
- Copyright holder:
- Kvasznicza, I
- Copyright date:
- 2014
If you are the owner of this record, you can report an update to it here: Report update to this record