Evolutionary Biology

, 36:416

Searching for Modular Structure in Complex Phenotypes: Inferences from Network Theory


    • División Antropología, Museo de La PlataUniversidad Nacional de La Plata, CONICET
    • Departamento de Biologia Animal, IBUniversidade Estadual de Campinas
  • Marcus A. M. de Aguiar
    • Departamento de Física da Matéria CondensadaIFGW, Universidade Estadual de Campinas
  • Paulo R. GuimarãesJr
    • Departamento de EcologiaInstituto de Biociências, Universidade de São Paulo
  • Sergio F. dos Reis
    • Departamento de Biologia Animal, IBUniversidade Estadual de Campinas
Tools and Techniques

DOI: 10.1007/s11692-009-9074-7

Cite this article as:
Ivan Perez, S., de Aguiar, M.A.M., Guimarães, P.R. et al. Evol Biol (2009) 36: 416. doi:10.1007/s11692-009-9074-7


The notion of modularity has become a unifying principle to understand structural and functional aspects of biological organization at different levels of complexity. Recently, deciphering the modular organization of molecular systems has been greatly aided by network theory. Nevertheless, network theory is completely absent from the investigation of modularity of complex macroscopic phenotypes, a fundamental level of organization at which organisms experience and interact with the environment. Here, we used geometric descriptors of phenotypic variation to derive a network representation of a complex morphological structure, the mammalian mandible, in terms of nodes and links. Then, by integrating the network representation and description with random matrix theory, we uncovered a modular organization for the mammalian mandible, which deviates significantly from an equivalent random network. The modules revealed by the network analysis correspond to the four morphogenetic units recognized for the mammalian mandible on a developmental basis. Furthermore, these modules are known to be affected only by particular genes and are also functionally differentiated. This study shows that the powerful formalism of network theory can be applied to the discovery of modules in complex phenotypes and opens the possibility of an integrated approach to the study of modularity at all levels of organizational complexity.


Geometric morphometricsCorrelation networksVariational modularitySimulated annealingMammalian mandible

Copyright information

© Springer Science+Business Media, LLC 2009