Display Settings:

Format

Send to:

Choose Destination
See comment in PubMed Commons below
J Theor Biol. 2002 Sep 21;218(2):175-85.

Analysis of comparative data using generalized estimating equations.

Author information

  • 1Laboratoire de Paléontologie, Paléobiologie & Phylogénie, Institut des Sciences de l'Evolution, Université Montpellier II, F-34095, Montpellier Cédex 05, France. paradis@isem.univ-montp2.fr

Abstract

It is widely acknowledged that the analysis of comparative data from related species should be performed taking into account their phylogenetic relationships. We introduce a new method, based on the use of generalized estimating equations (GEE), for the analysis of comparative data. The principle is to incorporate, in the modelling process, a correlation matrix that specifies the dependence among observations. This matrix is obtained from the phylogenetic tree of the studied species. Using this approach, a variety of distributions (discrete or continuous) can be analysed using a generalized linear modelling framework, phylogenies with multichotomies can be analysed, and there is no need to estimate ancestral character state. A simulation study showed that the proposed approach has good statistical properties with a type-I error rate close to the nominal 5%, and statistical power to detect correlated evolution between two characters which increases with the strength of the correlation. The proposed approach performs well for the analysis of discrete characters. We illustrate our approach with some data on macro-ecological correlates in birds. Some extensions of the use of GEE are discussed.

PMID:
12381290
[PubMed - indexed for MEDLINE]
PubMed Commons home

PubMed Commons

0 comments
How to join PubMed Commons

    Supplemental Content

    Full text links

    Icon for Elsevier Science
    Loading ...
    Write to the Help Desk