Understanding the rift between update rules in Evolutionary Graph Theory: The intrinsic death rate drives star graphs from amplifying to suppressing natural selection

Max Dew · 2025

Read Understanding the rift between update rules in Evolutionary Graph Theory: The intrinsic death rate drives star graphs from amplifying to suppressing natural selection by Max Dew (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-15528.txt.

Introduction

Max Dew and co-authors explain why two update rules give split verdicts on the same family tree graph. The work dates to 2025. In one rule a star graph boosts the spread of a new trait. In the other rule the same graph holds that spread back.

The team uses a smooth-time model that holds both rules as edge cases. They show the base death rate is the knob that flips the star from boost to brake. If the rate tops a bound tied to graph size, the star brakes selection. As graphs grow, that bound shrinks toward zero. Big stars thus brake for almost any base death rate.

Catalog record

CREATORMax Dew
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE55 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2506-15528

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Understanding the rift between update rules in Evolutionary Graph Theory: The intrinsic death rate drives star graphs from amplifying to suppressing natural selection Max Dew Abstract Evolutionary graph theory is the study of evolutionary dynamics in structured populations. A well-known problem in evolutionary graph theory is that the spread of mutation (measured by fixation probability) is impacted by the graph type chosen and the update rule. For example, the star graph is an amplifier of natural selection under the birth-death with fitness on birth (Bd) update rule but a suppressor of natural selection under the death-birth with fitness on birth (dB) update rule. A continuous-time EGT model has been found to replicate Bd and dB results as special cases. Using this model, we show that changing the natural (intrinsic) death rate can cause a shift from Bd to dB dynamics. Assuming the mutant is advantageous, we show that if the natural death rate is greater than \frac{1}{\sqrt{N}} the star is a suppressor, where N is the number of nodes. As N\longrightarrow\infty , the natural death rate required to drive the star to a suppressor tends towards zero, so as the size of the graph increases, the star graph is likely to be suppressing for any non-zero natural death rate. \ast Corresponding author: [email protected] , Orchid ID: https://orcid.org/0009-0009-7661-7369 \dagger Orchid ID: https://orcid.org/0000-0002-8433-4010 1 Introduction Evolution describes the process by which species adapt and change over time. The study of evolution is not only of interest due to scientific curiosity, but also allows the development of new techniques to make evolutionary predictions or influence evolution (evolutionary control) [ 1 ] , for example, predicting the spread of infectious diseases [ 2 ] . A key question in the study of evolution is that if a single mutant arises in a population of identical residents, what is the probability that the mutant takes over the resident population (known as fixation probability [ 3 , 4 , 5 , 6 ] )? One model to describe this kind of evolutionary process is the Moran process , first introduced in 1958 by Moran (1958) [ 7 ] and further developed in Moran (1962) [ 8 ] . The Moran process considers a finite haploid population consisting of two types of individuals, mutants and residents, who experience birth or death events [ 7 ] . The population is assumed to be homogeneous or well-mixed - every individual interacts with every other individual equally [ 9 ] . However, real populations are not necessarily well-mixed but rather heterogeneous [ 9 ] . For example, some microbiology experiments specifically study heterogeneous populations where the microbe populations are not shaken or well aerated, resulting in substructures [ 10 ] . While it is assumed that shaken aerated cultures are well-mixed [ 10 ] , some research has shown that this might not be the case [ 11 ] . Nowak et al. (2005) [ 12 ] extended the Moran process by considering heterogeneous populations (populations with a structure) represented by a graph [ 6 ] , creating a new field known as Evolutionary Graph Theory (EGT). The individuals exist on the nodes (or vertices) [ 5 , 6 ] , one individual per node [ 13 ] , and the edges represent an interaction between individuals [ 14 , 9 , 15 ] . With the introduction of heterogeneity (or structure) in the population, how does this impact the spread of mutation within this population? By comparing how the fixation probability changes relative to that of a homogeneous population, it is possible to see if the structure amplifies or suppresses the spread of mutation [ 4 , 9 ] . Nowak et al. (2005) [ 12 ] provided some examples, showing that the star graph is an amplifier of selection (increases spread of mutation), the burst is a suppressor of selection (decreases the spread of mutation) and the complete graph provides the same result as a homogeneous population from the Moran model. A problem that has arisen when determining the impact of population structure is that changing the update rule, i.e. the mechanism by which birth and death take place, can lead to qualitative differences on whether a certain graph is an amplifier or suppressor [ 16 ] . For example, for an update of birth-death with fitness on birth (Bd) [ 12 ] a star is an amplifier, but for death-birth with fitness on birth (dB) [ 17 ] , a star is a suppressor [ 15 ] . The aim of this project is to understand the biological drivers behind this qualitative change. We will discuss the known problem of the update rules with the classical EGT model, which we will refer to as discrete-time EGT model. We will then discuss a biologically motivated eco-evolutionary model [ 18 ] that is able to recreate the results from the discrete-time EGT model by suppressing ecological dynamics [ 18 ] , which we will refer to as continuous-time EGT model. Using the continuous-time EGT model, we will show that it is the natural death rate of individuals that drives the change from an amplifier to a suppressor. Focusing on the star graph as a well-known amplifier in Bd, we find an upper bound for this change in star graphs and show that, for very large systems, if the natural death rate is not zero, a star would be a suppressor. 2 Methods 2.1 Graphs Within evolutionary graph theory, we consider graphs defined by a right stochastic matrix W , where the element of the matrix w_{ij} [ 5 ] is the weight of the edge from node j to node i [ 19 ] , representing the strength of interaction between the two nodes [ 14 , 9 , 15 ] , defined as [ 5 ] W=\begin{cases}w_{ij}&\text{if there is an edge between }j\text{ and }i,\\ 0&\text{otherwise.}\\ \end{cases} (1) The nodes usually represent individuals [ 14 ] , but can also represent patches [ 20 ] , similar to meta-population models [ 16 , 18 , 21 ] . An undirected graph is where the edges are bidirectional [ 15 ] ( w_{ij}>0\text{ and }w_{ji}>0 [ 22 ] ) and unweighted if all edges leaving a node have equal weight [ 19 ] . The graphs contain self-loops if the node is connected to itself [ 15 ] ( w_{ii}\neq 0 [ 22 ] ). Only undirected unweighted graphs with self-loops are considered in this paper. 2.1.1 Complete graph A complete graph is a representation of homogeneous [ 9 ] or well-mixed populations [ 4 , 15 ] , and is defined as a fully connected graph [ 23 ] , where all the nodes are connected to every other node with equal weights. [ 9 ] The elements w_{ij} for a complete graph of N nodes with self-loops are defined as [ 18 ] w_{ij}=1/N\text{ for all $i$ and $j$.} (2) 2.1.2 Star graph A star graph with self-loops consists of two kinds of nodes - a central node, which connects to every node (including itself), and a leaf node, which connects only to the central node and itself [ 18 , 4 ] . The elements w_{ij}

Read the complete text