Fig. 7. SAGE method. (See text for details.)
tag counts from different experiments are directly comparable. Researchers from different labs can exchange SAGE libraries and compare experimental results (58). The SAGE method is often used in situations in which expression arrays could also be considered, comparing the expression patterns between tumor and normal cells (59). In fact, SAGE along with the Northern blot has been used as a validation of the microarray technology (60). SAGE offers some advantages over microar-rays in the discovery of novel genes, however. Although expression arrays usually require that a DNA sequence be known to be included on the array, SAGE samples all transcribed DNA. Limiting its broad use, however, is that compared to microarrays, the preparation of a SAGE library is laborious and expensive.
7.2. DNA-BASED GENOMIC ASSAYS: CGH, ARRAY CGH, AND SNP ARRAYS Thus, far we have used RNA quantity, as measured by RNA expression or cDNA fluorescence, to describe the activity of genes across the genome. There are also techniques reliant on genomic DNA to measure gene activity, useful when DNA is the only material available. Whereas RNA can only be harvested from fresh tissue under ideal circumstances, DNA is relatively stable and can be isolated from paraffin-embedded tissue. Although RNA is tightly regulated and variations in its abundance clearly relate to changes in cellular behavior, DNA is present in consistent concentration in all diploid cells. DNA-based genomic assays exploit the pathologic exceptions to this rule, in which increased or decreased amounts of DNA are observable. When associated with disease, alterations in DNA copy number, in turn, likely correspond to changes in RNA through the phenomenon of gene "dose." Cancer provides the best examples, in which many measurable genetic abnormalities play a role in tumor genesis, including translocations, chromosome loss (partial or whole chromosome), and chromosome gain. The ability to survey across the genome for loss of DNA regions containing tumor suppressor genes or gain of DNA regions with onco-genes is of great interest to both clinicians and researchers.
An early DNA based genomic assays called comparative genomic hybridization (CGH) was first described in 1992 in the journal Science (61). The technique involves first labeling reference and test DNA samples with different fluorescent dyes (red and green) and then hybridizing to normal metaphase chromosomes. The relative fluorescence between test and reference sample at different loci along the chromosome is proportional to the relative number of DNA copies. In this way, amplifications of whole chromosomes or portions of the chromosome are reliably detected, although the resolution of these regions is low, in the range of 20 Mb. Once a locus of interest has been identified, investigators can attempt to identify genes within it by fluorescent in situ hybridization (FISH). As the resolution of the method is quite coarse, gene identification through CGH can be a laborious process.
In 1998, Pinkel et al. described a microarray-based technique that improved the resolution of CGH dramatically from 20 Mb to 40 kb (62). Instead of using metaphase chromosomes as the probes, the authors used a library of genomic DNA, each sequence mapped to a known location on a chromosome. The genomic DNA probes are spotted on a glass slide, as were the cDNA probes from the microarray example in Section 3.2. Also similar is the process of labeling test and reference DNA with different fluorescent probes, hybridizing against the microarray, and scanning for relative fluorescence. In this manner, individual chromosomes and in fact the entire genome can be evaluated for areas of amplification or deletion with a much finer resolution than earlier. As with CGH, a DNA region of importance is highlighted by array CGH, not individual genes. At this resolution, however, using genomic sequences, an investigator can narrow candidate genes to a relatively small number for further investigation. Refinements continue to be made to this technique to increase the resolution of the assay (63).
An alternate DNA microarray-based genomic approach is the SNP (pronounced "snip") array. The most frequent variation in the human genome is a single base substitution called a single-nucleotide polymorphism (SNP). To be considered a SNP, and not a mutation, an event frequency should be >1%, although many occur with much higher frequency. As expected, most SNPs occur in nontranscribed portions of the genome, where they would be thought to have little biologic significance. Although these nontranscribed SNPs have no functional significance, they provide a useful marker for genetic investigation in the following manner. A somatic cell is sampled to determine the germline genotype at a SNP locus. If the cell is heterozygous for the SNP, then the site is potentially informa tive in evaluating a tumor for DNA loss. In other words, if the somatic cell is heterozygous for a SNP and the tumor cell is homozygous, this suggests that one chromosome or a part of one chromosome containing the missing allele has been lost. Loss of genetic material in this way is termed "loss of heterozygosity" (LOH). Homozygous SNP loci in the somatic cell provide no information on LOH in the tumor cell.
Affymetrix has produced versions of its oligonucleotide array that uses these principles to identify LOH in tumor cells (64,65). The current version contains probe sets to genotype approx 10,000 human SNPs selected for their high frequency of heterozygosity and even coverage of the genome. The chips are prepared in exactly the same manner as those described in Section 3.5, although the probe design strategy is different. Probes are again 25 bp in length, with the base in position 13 representing the SNP. One probe for each possible substitution at the SNP site is included on the chip as well as probes for flanking sequences of DNA. In this way normal somatic cells can be assayed to determine the loci at which an individual is heterozygous. Tumor cells are assayed and the results compared. Heterozygous loci in the somatic cell that register as homozygous in the tumor cell indicate areas of interest in tumor biology, especially when they occur in multiple samples. The pattern of LOH itself might be useful for identifying tumor subtypes. Areas of LOH can highlight the location of tumor related genes, such as tumor suppressor genes or oncogenes.
7.3. OTHER ARRAYS Having described two examples each of DNA and RNA genomics assays, we acknowledge that many other similar technologies are in various stages of development. We have focused only on those with the most immediate clinical applications and that are perhaps the most mature. Additionally, we have omitted a wide variety of microarray-based technologies that are not, strictly speaking, genomics techniques. Examples of these include the tissue microarray, in which small tissue samples from many sources are assembled and displayed on a single glass slide (66). Tissue arrays provide an efficient method for measuring many samples for features such as protein or DNA mutations. The slide can be stained for the characteristic and scanned for its prevalence across the many samples on the array (67). Other powerful DNA microarray technologies have been developed for the purposes of mutation detection, sequencing, genotyp-ing, and polymorphism detection. These are, again, not generally genomics tools and we refer the reader to the original publications and reviews (68-71).
Was this article helpful?
Are You Prepared For Your First Baby? Endlessly Searching For Advice and Tips On What To Expect? Then You've Landed At The Right Place With All The Answers! Are you expecting? Is the time getting closer to giving birth to your first baby? So many mothers to be are completely unprepared for motherhood and the arrival of a little one, but stress not, we have all the answers you need!