Regulatory network construction in Arabidopsis by using genome-wide gene expression quantitative trait loci

doi:10.1073/pnas.0610429104

Regulatory network construction in Arabidopsis by using genome-wide gene expression quantitative trait loci

Laboratories of *Genetics and
^†Plant Physiology, Wageningen University, Arboretumlaan 4, NL-6703 BD Wageningen, The Netherlands;
^§Groningen Bioinformatics Centre, Groningen Biomolecular Sciences and Biotechnology Institute, University of Groningen, Kerklaan 30 NL-9751 NN Haren, The Netherlands;
^¶Molecular Genetics Group, Department of Biology, Utrecht University, Padualaan 8, NL-3584 CH Utrecht, The Netherlands;
^‖Plant Ecophysiology, Institute of Environmental Biology, Utrecht University, Sorbonnelaan 16, NL-3584 CA Utrecht, The Netherlands; and
**Max Planck Institute for Plant Breeding Research, Carl-von-Linné-Weg 10, 50829, Cologne, Germany

Contributed by Maarten Koornneef, November 24, 2006 (received for review September 29, 2006)

Abstract

Accessions of a plant species can show considerable genetic differences that are analyzed effectively by using recombinant inbred line (RIL) populations. Here we describe the results of genome-wide expression variation analysis in an RIL population of Arabidopsis thaliana. For many genes, variation in expression could be explained by expression quantitative trait loci (eQTLs). The nature and consequences of this variation are discussed based on additional genetic parameters, such as heritability and transgression and by examining the genomic position of eQTLs versus gene position, polymorphism frequency, and gene ontology. Furthermore, we developed an approach for genetic regulatory network construction by combining eQTL mapping and regulator candidate gene selection. The power of our method was shown in a case study of genes associated with flowering time, a well studied regulatory network in Arabidopsis. Results that revealed clusters of coregulated genes and their most likely regulators were in agreement with published data, and unknown relationships could be predicted.

natural variation

Footnotes

^‡To whom correspondence may be addressed. E-mail: joost.keurentjes{at}wur.nl or koornnee{at}mpiz-koeln.mpg.de
Author contributions: J.J.B.K., J.F., I.R.T. contributed equally to this work; J.J.B.K., G.v.d.A., A.J.M.P., D.V., M.K., and R.C.J. designed research; J.J.B.K., I.R.T., and J.M.G. performed research; J.J.B.K., J.F., I.R.T., L.B.S., and R.C.J. analyzed data; and J.J.B.K., J.F., I.R.T., G.v.d.A., A.J.M.P., D.V., M.K., and R.C.J. wrote the paper.
The authors declare no conflict of interest.
This article contains supporting information online at www.pnas.org/cgi/content/full/0610429104/DC1.
Abbreviations:

QTL,

quantitative trait locus;

eQTL,

expression quantitative trait locus;

iGA,

iterative group analysis;

RIL,

recombinant inbred line;

Ler,

Landsberg erecta;

Cvi,

Cape Verde Islands;

FDR,

false-discovery rate;

Col,

Columbia;

PC,

possibility of change.