1. ¹UMR Biodiversité, Gènes & Ecosystèmes, INRA UMR BIOGECO, 69 route d'Arcachon, 33612 Cestas Cedex, France
2. ²CEH – Edinburgh, Bush Estate, Penicuik, Midlothian EH26 OQB, UK
3. ³IPBO, Ghent University, Ledeganckstraat 35, Ghent 9000, Belgium
4. ⁴Department of Molecular Biotechnology, Ghent University, Coupure links 653, Ghent 9000, Belgium
5. ⁵INPA, Avenida A Araujo, 2936, Manaus 69083-300, Brazil
6. ⁶School of Integrative Biology, University of Queensland, Brisbane, Queensland 4072, Australia
7. ⁷UFRJ Laboratório de Genética Molecular Vegetal, Departamento de Genética, IB, Universidade Federal do Rio de Janeiro (UFRJ), Rio de Janeiro, Brazil
8. ⁸Departamento de Bioquímica, Universidade Federal do Rio Grande do Sul, UFRGS, Porto Alegre, Brazil
9. ⁹Centro Agrico Tropical de Investigacia Ensena (CATIE), Cartago, Turrialba 7170, Costa Rica

Correspondence: A Kremer, UMR Biodiversité, Gènes & Ecosystèmes, INRA UMR BIOGECO, 69 route d'Arcachon, 33612 Cestas Cedex, France. E-mail: kremer@pierroton.inra.fr

Received 8 February 2005; Accepted 21 July 2005; Published online 17 August 2005.

Abstract

Since no universal codominant markers are currently available, dominant genetic markers, such as amplified fragment length polymorphism (AFLP), are valuable tools for assessing genetic diversity in tropical trees. However, the measurement of genetic diversity (H) with dominant markers depends on the frequency of null homozygotes (Q) and the fixation index (F) of populations. While Q can be estimated for AFLP loci, F is less accessible. Through a modelling approach, we show that the monolocus estimation of genetic diversity is strongly dependent on the value of F, but that the multilocus diversity estimate is surprisingly robust to variations in F. The robustness of the estimate is due to a mechanistic effect of compensation between negative and positive biases of H by different AFLP loci exhibiting contrasting frequency profiles of Q. The robustness was tested across contrasting theoretical frequency profiles of Q and verified for 10 neotropical species. Practical recommendations for the implementation of this analytical method are given for genetic surveys in tropical trees, where such markers are widely applied.