Email updates

Keep up to date with the latest news and content from BMC Genetics and BioMed Central.

Open Access Highly Accessed Research article

Association studies including genotype by environment interactions: prospects and limits

Abdoul-Aziz Saïdou123456*, Anne-Céline Thuillet1, Marie Couderc1, Cédric Mariac12 and Yves Vigouroux16*

Author Affiliations

1 Institut de Recherche pour le Développement, UMR DIAPC IRD/INRA/Université de Montpellier II/ Montpellier SupAgro, BP64501, 34394 Montpellier, France

2 Institut de Recherche pour le Développement, UMR DIAPC IRD/INRA/Université de Montpellier II/ Montpellier SupAgro, BP11416 Niamey, Niger

3 Université Abdou Moumouni, BP 11040 Niamey, Niger

4 Montpellier SupAgro, 2, place Pierre Viala, 34060 Montpellier, France

5 Current address: UMR CEFE, 1919 Route de Mende, Montpellier, France

6 Institut de Recherche pour le Développement, 911, avenue Agropolis, 34394 Montpellier, France

For all author emails, please log on.

BMC Genetics 2014, 15:3  doi:10.1186/1471-2156-15-3

Published: 6 January 2014



Association mapping studies offer great promise to identify polymorphisms associated with phenotypes and for understanding the genetic basis of quantitative trait variation. To date, almost all association mapping studies based on structured plant populations examined the main effects of genetic factors on the trait but did not deal with interactions between genetic factors and environment. In this paper, we propose a methodological prospect of mixed linear models to analyze genotype by environment interaction effects using association mapping designs. First, we simulated datasets to assess the power of linear mixed models to detect interaction effects. This simulation was based on two association panels composed of 90 inbreds (pearl millet) and 277 inbreds (maize).


Based on the simulation approach, we reported the impact of effect size, environmental variation, allele frequency, trait heritability, and sample size on the power to detect the main effects of genetic loci and diverse effect of interactions implying these loci. Interaction effects specified in the model included SNP by environment interaction, ancestry by environment interaction, SNP by ancestry interaction and three way interactions. The method was finally used on real datasets from field experiments conducted on the two considered panels. We showed two types of interactions effects contributing to genotype by environment interactions in maize: SNP by environment interaction and ancestry by environment interaction. This last interaction suggests differential response at the population level in function of the environment.


Our results suggested the suitability of mixed models for the detection of diverse interaction effects. The need of samples larger than that commonly used in current plant association studies is strongly emphasized to ensure rigorous model selection and powerful interaction assessment. The use of ancestry interaction component brought valuable information complementary to other available approaches.

Association study; G × E; Power simulation; Model selection; REML; PHYC; Vgt1