A comprehensive assessment of the transcriptome of cork oak (Quercus suber) through EST sequencing

Abstract

Background: Cork oak (Quercus suber) is one of the rare trees with the ability to produce cork, a material widely used to make wine bottle stoppers, flooring and insulation materials, among many other uses. The molecular mechanisms of cork formation are still poorly understood, in great part due to the difficulty in studying a species with a long life-cycle and for which there is scarce molecular/genomic information. Cork oak forests are of great ecological importance and represent a major economic and social resource in Southern Europe and Northern Africa. However, global warming is threatening the cork oak forests by imposing thermal, hydric and many types of novel biotic stresses. Despite the economic and social value of the Q. suber species, few genomic resources have been developed, useful for biotechnological applications and improved forest management.

Results: We generated in excess of 7 million sequence reads, by pyrosequencing 21 normalized cDNA libraries derived from multiple Q. suber tissues and organs, developmental stages and physiological conditions. We deployed a stringent sequence processing and assembly pipeline that resulted in the identification of ~159,000 unigenes. These were annotated according to their similarity to known plant genes, to known Interpro domains, GO classes and E.C. numbers. The phylogenetic extent of this ESTs set was investigated, and we found that cork oak revealed a significant new gene space that is not covered by other model species or EST sequencing projects. The raw data, as well as the full annotated assembly, are now available to the community in a dedicated web portal at http://www.corkoakdb.org.

Conclusions: This genomic resource represents the first trancriptome study in a cork producing species. It can be explored to develop new tools and approaches to understand stress responses and developmental processes in forest trees, as well as the molecular cascades underlying cork differentiation and disease response.

Figure 1

Schematic representation of the bioinformatics pipeline, indicating the software used at each step.

Figure 2

Assembly and predicted peptide statistics. (A) Unigene length distribution after multi-library assembly. There are 12 additional unigenes longer than 4600 bases, not shown on the plot, with the longest one being 9189 bases. (B) Unigene coverage (reads per unigene). (C) Serial clustering of predicted proteins based on the cork oak unigenes, and of the predicted proteins from the genomes of two model plant species.

Figure 3

Gene Ontology classification of nuclear unigenes. Classification was performed using CateGOrizer, counting single occurrences and the Generic GO Slim [25]. Percentages are shown down to 3% only, and the functional classes are ordered by frequency.

Figure 4

Distribution of annotation classes in the cork oak translated unigenes.

Figure 5

Unique Interpro domains assigned to the Q. suber unigenes and two other transcriptomes for Q. robur and Castanea mollissima , as well as for species with completely sequenced genomes A. thaliana, P. trichocarpa and P. persica .

Figure 6

Number of the cork oak’s predicted peptides unique BLAST hits in other plant genomes.

Figure 7

Overlap between the cork oak unigenes (brown) and the unigenes of the red oak, English oak and Chinese chestnut. Numbers represent homologues defined at a e < 10^-5 cut off, and in parentheses at e < 10^-2.

Figure 8

CorkOakdb.org. Screenshot of the top part of the gene view.