Registering the evolutionary history in individual-based models of speciation

TL;DR

This paper introduces two algorithms to record evolutionary histories in individual-based models of speciation, enabling the construction of genealogies and phylogenies to better understand macroevolutionary patterns.

Contribution

The paper presents novel algorithms for capturing detailed evolutionary histories in individual-based models, improving the accuracy of phylogenetic reconstructions.

Findings

MRCAT trees are closer to true phylogenies than distance-based trees.

Both algorithms reveal differences between inferred and true evolutionary histories.

Algorithms facilitate analysis of speciation and extinction processes in models.

Abstract

Understanding the emergence of biodiversity patterns in nature is a central problem in biology. Theoretical models of speciation have addressed this question in the macroecological scale, but little has been investigated in the macroevolutionary context. Knowledge of the evolutionary history allows the study of patterns underlying the processes considered in these models, revealing their signatures and the role of speciation and extinction in shaping macroevolutionary patterns. In this paper we introduce two algorithms to record the evolutionary history of populations in individual-based models of speciation, from which genealogies and phylogenies can be constructed. The first algorithm relies on saving ancestral-descendant relationships, generating a matrix that contains the times to the most recent common ancestor between all pairs of individuals at every generation (the Most Recent Common Ancestor Time matrix, MRCAT). The second algorithm directly records all speciation and extinction events throughout the evolutionary process, generating a matrix with the true phylogeny of species (the Sequential Speciation and Extinction Events, SSEE). We illustrate the use of these algorithms in a spatially explicit individual-based model of speciation. We compare the trees generated via MRCAT and SSEE algorithms with trees inferred by methods that use only genetic distance among extant species, commonly used in empirical studies and applied here to simulated genetic data. Comparisons between tress are performed with metrics describing the overall topology, branch length distribution and imbalance of trees. We observe that both MRCAT and distance-based trees differ from the true phylogeny, with the first being closer to the true tree than the second.