On a Genetics Model of Moran

Samuel Karlin, James McGregor

Mathematical Proceedings of the Cambridge Philosophical Society · 1962 · 98 citations · 13 references

Concepts

TL;DR

Stochastic genetics models, such as those by Wright and Feller, study gene frequency fluctuations under mutation and selection, often using a simple model with a fixed number N of gametes of two types a and A. The Moran model defines a Markov chain X(n) where state j denotes j a‑gametes, each next‑generation gamete is independently sampled from the current population with mutation probabilities γ1 (a→A) and γ2 (A→a), yielding transition probabilities proportional to the current gamete fractions.

Abstract

Wright (14), Feller (2) and others have proposed various stochastic models of genetics to study the fluctuations of gene frequency under the influence of mutation and selection. One of their simplest models has the following structure. There are a fixed number N of gametes each of which may be of two types a or A . The process X(n) , n = 0, 1, 2, … which is assumed to have stationary transitions, is said to be in state j when there are j gametes of type a , and N − j of type A . Let γ 1 denote the probability that immediately after formation an a gamete mutates into an A gamete and let γ 2 denote the probability of an A gamete mutating into an a gamete. Each of the gametes of the next generation is independently formed by making a random selection from the gametes of the present generation. The probability of a particular gamete in the next generation being of type a is then and of type A where j represents the state of the process. Roughly speaking, the chance of a mating resulting in a gamete of a prescribed kind for the next generation is proportional to the fraction of these gametes present in this generation allowing for mutation effects.

References

13