A general evolutionary model for the emergence of novel characters from serial homologs

Abstract

Repeated structures are widespread across multicellular organisms, such as vertebrae, segments, and cell types. These structures, also known as serial homologs, share ancestral states and developmental underpinnings yet also provide the materials for novel adaptive phenotypes. It remains largely unclear why some repeated structures diverge quickly, while others remain constrained. One reason for this uncertainty is the lack of a generalized model for the evolution of repeated structures under different scenarios that links developmental, genetic, and selective constraints to expected and observed patterns of evolution. Here, we introduce a model that incorporates key structural features of gene regulatory networks and selection and investigate how responses to multivariate selection depend on developmental constraints. We show structural features of developmental networks determine when repeats can respond independently to selection and when divergence is limited. Simulations recover broad expectations of phenotypic evolution under directional selection inferred from empirical data. We further show that, in the face of fluctuating selection, strong developmental constraints lead to reduced fitness over time and attenuated fitness fluctuations. Together, our results provide general insights into the principles of evolution of repeated structures and offer a modeling framework for the evolution of a broad range of key phenotypic characters.