Y. Bar-Yam
Abstract:A historical dispute in the conceptual underpinnings of evolution is the validity of the gene centered view of evolution. \cite{dawkins,lewontin} We transcend this debate by formalizing the gene centered view as a dynamic version of the mean field approximation. This establishes the conditions under which it is applicable and when it is not. In particular, it breaks down for trait divergence which corresponds to symmetry breaking in evolving populations.
What problem does this paper attempt to address?
The core problem that this paper attempts to solve is about the validity and scope of application of the "gene - centric view" in evolutionary theory. Specifically, the author formalizes the gene - centric view as a dynamic version of the mean - field approximation to determine under which conditions this view is applicable and under which conditions it is not.
### Main problems:
1. **Validity of the gene - centric view**: Traditionally, there are two main perspectives in evolutionary theory: one is the view based on individual selection, and the other is the view based on gene selection. The gene - centric view holds that in sexually reproducing organisms, genes are passed on from one generation to the next as indivisible units, and there is competition between different versions of genes (i.e., alleles), not between individuals.
2. **Conditions of application and limitations**: The author attempts to clearly answer under what conditions the gene - centric view can be an effective substitute for individual competition? Especially in the process of sexual reproduction, since offspring may be very different from their parents, this complicates simple evolutionary views. Therefore, the author introduces a dynamic form of the mean - field approximation to analyze this problem.
### Specific problems:
- **Formalization of selection and reproduction steps**: The author formalizes the two key steps in evolution - selection and reproduction - into mathematical equations:
\[
\{N(s;t)\} = R[\{N'(s;t - 1)\}]
\]
\[
\{N'(s;t)\} = D[\{N(s;t)\}]
\]
where \( N(s;t) \) represents the number of individuals with a specific genome \( s \) at time \( t \), and \( N'(s;t) \) represents the number of selected individuals at the same time point.
- **Application of the mean - field approximation**: The author introduces the mean - field approximation, assuming that the behavior of each gene can be regarded as being in an average environment formed by other genes. The formula is expressed as:
\[
P(s_1,\ldots, s_N)=\prod_i P(s_i)
\]
This means that the probability distribution of each gene is independent, ignoring the correlation between genes.
- **Dependence of fitness**: The author points out that the effective fitness of each allele depends on the distribution of other alleles in the population, which leads to the time - dependence of fitness. For example, the probability of victory for a right - handed rower in a predominantly left - handed population is different from that in a predominantly right - handed population.
- **Symmetry Breaking**: When trait divergence occurs in the population, the mean - field approximation will fail. Trait divergence may be caused by correlations in the reproduction process (such as spatial separation or kin selection), which will cause genes to no longer be completely mixed.
### Conclusion:
The gene - centric view is applicable to those populations in which sexual reproduction results in complete allele mixing, but on the premise that fitness must be understood relative to the current gene pool. However, for structured populations (such as species with local mating areas), or when trait divergence occurs, the gene - centric view is not applicable. These problems are very important for understanding ecological diversity, sympatric speciation, etc.
In this way, the author not only solves the problem of the applicability of the gene - centric view, but also provides a new perspective for understanding complex phenomena in evolutionary biology.