Abstract
Gene conversion is a genetic mechanism by which one gene is `copied and pasted' onto another gene, where the direction can be biased between the different types. In this paper, a stochastic model for biased gene conversion within a d-unlinked multigene family and its diffusion approximation are developed for a finite Moran population. A connection with a d-island model is made. A formula for the fixation probability in the absence of mutation is given. A two-timescale argument is applied in the case of the strong conversion limit. The dual process is generally shown to be a biased voter model, which generates an ancestral bias graph for a given sample. An importance sampling algorithm for computing the likelihood of the sample is deduced.
Citation
Shuhei Mano. "Ancestral graph with bias in gene conversion." J. Appl. Probab. 50 (1) 239 - 255, March 2013. https://doi.org/10.1239/jap/1363784436
Information