While ESTs have often been sequenced to access both gene sequence from a given genome and to derive an approximate expression level for the sequenced transcripts on the basis of the underlying sequence redundancy, new technologies have come to the forefront that allow for the parallel investigation of both relative and comparative expression levels within series of experiments.
The northern blot (Alwine et al. 1977) has become an indispensable method for the quantification of RNA molecules that have been resolved on a gel and immobilised onto a solid substrate. A labelled DNA 'probe' is hybridised to these immobilised RNAs and the resulting quantification of label provides a view of transcript abundance. A reverse procedure was demonstrated using 45 Arabidopsis genes. A probe to each gene was immobilised on a glass slide and free-labelled RNA extracts were hybridised to the first arrays. Quantification of label from each gene allowed for the parallel investigation of expression across the whole set (Schena et al. 1995). This simple demonstration of an array technology has naturally evolved and now 48,000 probes may be comfortably fitted on a single glass slide, hundreds-of-thousands of features may be placed on other more proprietary platforms such as those from companies including Illumina, Agilent, Affymetrix or Nimblegen.
An EST sequence stems from a cDNA clone. The cDNA insert that has been sequenced may be mechanically applied to a treated glass-slide to create a cDNA array. The cDNA may be spotted prior to sequencing such that a large number of candidate genes that are differentially expressed following a particular treatment may be identified and then sequenced e.g. (Lim 2005). Alternatively the cDNAs may be arrayed after the sequencing step so that in addition to the selection of new candidate genes, already known sequences may be investigated for quantitative differences within an experiment. The application of cDNA arrays in contemporary crop biology has become extremely widespread and brief literature review identifies publications relating to Arabidopsis (Kim and von Arnim 2006) (Oono et al. 2006), cotton (Shi et al. 2006), medicago (Tesfaye et al. 2006), sorghum (Buchanan et al. 2005), wild rice (Kim et al. 2005), potato (Rensink et al. 2005) (Schmidt et al. 2005), the genus Senecio (Hegarty et al. 2005), poplar (Taylor et al. 2005), cassava (Lopez et al. 2005), citrus (Forment et al. 2005), gerbera (Laitinen et al. 2005), eucalyptus (Duplessis et al. 2005), Brassica oleracaea (Soeda et al. 2005), strawberry (Aharoni et al. 2004), tobacco (Matsuoka et al. 2004) and pine (Egertsdotter et al. 2004).
The continued popularity of cDNA microarrays is in part driven by the relative inexpensiveness of physically arraying small aliquots of DNA solution onto a glass slide. Since no a priori knowledge as to the content and structure of the genes expressed within a tissue is needed, cDNA arrays are inexpensive to set-up and are amenable to customisation (groups of target genes may be easily added to the array). The array construction process can be further simplified by arraying the DNA solution onto nylon filters yielding 'macro-arrays'. Macroarrays have a much lower feature density and a typical filter may contain only a few thousand features at most. Such arrays continue to be used within genomics research e.g. (Beldade et al. 2006; Derory et al. 2006; Jia et al. 2006; Nakano et al. 2006; Puthoff and Smigocki 2007), but the availability of many academic and commercial service providers have driven the popularity of both cDNA and oligonucleotide microarrays. Another critical consideration of the macroarray technology is the physical size of the array (tens of square centimetres) and the necessary volumes of hybridisation solutions that are required. Macroarrays are therefore unsuited to some of the more contemporary and sensitive techniques within gene expression profiling. It might be argued that macroarrays are best suited to pilot projects, small numbers of candidate genes, or to preselected clusters of pre-classified genes.
Oligonucleotide arrays instead of being reliant upon a cloned and amplified cDNA molecule use instead the sequence to select for long DNA oligonucleotide sequences that may be between 25 and 80 nucleotides long. These oligonucleotides may be synthesised and mechanically arrayed onto a glass slide, they may be synthesised on micro-beads and arrayed or may be synthesised directly on an array as exemplified by the Affymetrix photolithography process. Commerical oligonucleotide arrays such as those provided by Affymetrix have been widely adopted by the research community since they may provide greater reproducibility and sensitivity than cDNA arrays.
Since the manufacture of oligonucleotide arrays requires access to deep quality sequence information this has recently been restricted to the model organisms. However, the demands of the crop research community has been such that oligonucleotide arrays are commercially available on the Affymetrix platform for Arabidopsis thaliana, Hordeum vulgare, Zea mays, Medicago truncatula, Oryza sativa, Populus trichocarpa, Glycine max, Saccharum officinarium, Lycopersicon esculentum, Triticum aestivum and Vitis vinifera. Other companies such as Illumina and Nimblegen have methods for probe design and optimisation and a ready to prepare oligonucleotide arrays to suit the needs of the crop research community. It would seem that with today's broad and deep EST collections that meaningful oligonucleotide arrays could be synthesised to address many questions and to identify candidate genes involved in many biological processes.
The popularity of microarrays as a fundamental technology to view differential gene expression and as a bridge-technology into the field of system biology or functional genomics is clear (Allison et al. 2006). It has been argued that if significant EST, or other genomic resource, exist then eventually a microarray will be produced (Richmond and Somerville 2000). The varied crop plants for which cDNA and oligonucleotide arrays are already available show that this argument is indeed completely true; furthermore, there are cases where ESTs have undoubtedly been sequenced as a step in the construction of a microarray.
The roles of microarrays within plant genomics continue to diversify. The development of techniques for array-based single nucleotide polymorphism (SNP) classification and the concomitant genome-scale genotyping and haplotyping strategies are opening exciting new developments for the plant breeders (Borevitz 2006). Array-based SNP technologies have already been demonstrated in at least potato (Rickert et al. 2005) and rice (Shirasawa et al. 2006) and will undoubtedly be described in many more species. This exciting new direction is perhaps limited by our ability to define the starting SNPs rather than in their subsequent detection (Chevreux et al. 2004).
Was this article helpful?