Submitted to: Plant Journal
Publication Type: Peer reviewed journal
Publication Acceptance Date: 8/29/2007
Publication Date: 1/2/2008
Citation: Potokina, E., Druka, A., Luo, Z.W., Wise, R.P., Waugh, R., Kearsey, M.J. 2008. Gene Expression Quantitative Trait Locus Analysis of 16,000 Barley Genes Reveals a Complex Pattern of Genome-wide Transcriptional Regulation. Plant Journal. 53(1):90-101. Interpretive Summary: The analysis of traits in crop plants has relied heavily on the methodologies of quantitative genetics and analysis of Quantitative Trait Loci (QTL). Generally, such crop species have un-sequenced genomes and require syntenic information from sequenced model species to interpret molecular marker information. However, the utilization of high-content parallel expression analysis enables the transcriptome to be viewed as an additional phenotype in QTL analysis for crops. To date, most developments in these areas have been carried out in sequenced model organisms, such as yeast or Arabidopsis. High density oligonucleotide arrays provide a powerful tool for genetical genomics studies in crop plant species allowing the transcriptome profiles of thousands of genes to be performed for every line of a mapping population. This allows the expression level for each gene on the array to be treated as a quantitative trait that can be scored for all progeny lines and analyzed by classical QTL methodologies. Species with fully sequenced genomes provide the physical genetic information to produce whole genome coverage of all known genes on expression arrays, while in other species it is necessary to rely on sequenced ESTs (Expressed Sequence Tags) from selected tissues. This article reports a genetical genomics approach to ‘whole plant’ material of the well studied Steptoe (St) x Morex (Mx) segregating population using the Affymetrix Barley1 GeneChip. The design of the Affymetrix Barley1 GeneChip allows two goals to be achieved simultaneously: i) to extract polymorphic markers from the RNA profiling data, and ii) to measure the expression level for each gene on a chip across all progeny lines in a mapping population, and thus identify gene expression QTL (eQTLs). This approach is of particular value when the traits are of commercial importance and are the subject of selective improvement for food, energy, and other materials. Based on analysis of transcript level variation of approximately 22 thousand genes across 139 DH (doubled-haploid) lines of the St/Mx cross, we applied a simple and efficient algorithm to identify a very large number of polymorphic markers (Transcript Derived Markers - TDMs) in cRNA profiling data. With this set of TDM markers, we constructed a genetic map and used that map for the subsequent genome-wide eQTL analysis of approximately 16 thousand genes. Impact: The availability of such an approach has wide value in genetical analysis of crop plants and farm animals and also in complementing information available in sequenced model organisms, thus translating information from model systems to crops of significant value.
Technical Abstract: Transcript abundance data from cRNA hybridizations to Affymetrix microarrays can be used for simultaneous marker development and genome-wide eQTL (expression Quantitative Trait Loci) analysis of crops. We have shown that it is easily possible to use the information from Affymetrix expression arrays used to profile individuals from a segregating population to accurately identify robust polymorphic molecular genetic markers. We identified over 2000 genetic polymorphisms (TDMs - Transcript Derived Markers) from an experiment involving two commercial varieties of barley [Steptoe (St) and Morex (Mx)] and their doubled-haploid (DH) progeny. We constructed a genetic map of the TDMs and used it for genome-wide eQTL analysis of approximately 16 thousand genes in the mapping population of 139 DH lines. We identified 23,738 significant gene expression QTL (eQTL) at a genome-wide significance (P < 0.05), affecting the expression of 12,987 genes. Over a third of these genes with expression variation have only one identified eQTL while the rest have 2-6, involving cis- and trans effects. Overall, more than half of the quantitatively controlled transcripts appear to be primarily regulated by cis-eQTLs in the St x Mx population. We show that, although there appear to be hotspots of eQTL across the seven chromosomes, many are associated with regions of low recombination such as the genetic centromeres and so have high gene density per cM. However, some chromosomal regions have a significant excess of eQTL over the number expected from gene density estimated by TDMs and genic SNP markers, and many of these are biased towards eQTL for which alleles from one particular parent increases expression level.