Background
Knowing the differences in gene expression levels among different tissues is of paramount interest in various areas of biomedical research. These include, but are by no means limited to, (i) comparisons of healthy and diseased tissue in order to understand malfunctions in regulation and identify genes essential for control, thus identifying target molecules for the development of novel therapeutics and (ii) the observation of changes in expression over time after addition of a drug to elucidate the mechanism of action of pharmaceuticals and predict their toxicology. Experiments in that realm furthermore help to understand the molecular basis of diseases, provide means for early diagnostics, and facilitate monitoring of therapy.
Examples for analog techniques representing mRNA expression patterns are subtractive cDNA libraries [1-8] and the Differential Display method and its derivatives [9-12]. Macro- and microarrays [13] are rapidly becoming the analog method of choice, mainly due to their high throughput in data production which allows complex biological questions to be addressed.
Digital methods for counting differences in gene expression levels provide specific advantages. With the expressed sequence tag (EST) approach [14,15] expression patterns are analyzed by sequencing many clones from cDNA libraries. Even limited sequence information on the cDNA 3'-end (tag sequences) permits unambiguous identification of the cDNA and the corresponding gene. The different frequencies of cDNAs in libraries derived from different sources give evidence of possible changes in gene expression. This approach provides accurate quantitative information and has a flexible degree of sensitivity which depends solely on the number of analyzed clones. However, this is very labor-intensive. In order to analyze the expression of low abundant cDNAs that represent mRNA in the range of one copy per mammalian cell, more than 100.000 colonies have to be analyzed per library. Improvements in sequencing automation and analysis such as capillary electrophoresis have speeded up the process considerably and recent developments such as a sequencing method based on real-time pyrophosphate [16], sequencing on microchips [17], and massive parallel signature sequencing on microbead arrays [18] also contribute to the speed and depth of EST gene expression analysis. Oligonucleotide fingerprinting [19] characterizes expressed genes via the hybridization of hundreds of synthetic oligonucleotides to cDNA that produces unique fingerprints of matching and non-matching oligonucleotides. Another approach to increase throughput is the serial analysis of gene expression (SAGE). Short defined cDNA sequences are initially prepared from mRNA which are then dimerized, multimerized, cloned, and sequenced [20]. SAGE accelerated the process of gene expression analysis more than an order of magnitude compared to the conventional EST analysis and still is compatible with most of the improvements of DNA sequencing mentioned above. However, after the identification of differentially expressed cDNA the full length gene has to be cloned starting from minor sequence information (12 bp tag) which hampers further functional analysis of the corresponding protein, a complex task that requires the complete coding sequence.
Each of these methods clearly has its specific advantages, and often different subsets of differentially expressed genes are identified employing a certain method. For example, cDNAs that are not easily amplified by means of PCR are inadequately represented in PCR-based methods. DADA was designed in order to overcome specific shortcomings of established differential expression analysis technologies used today, e.g. the need to sequence all of the examined genes or the involvement of PCR steps. In addition, further cloning of the complete coding sequence of the identified genes of interest is facilitated by ending up with rather long (e.g. in comparison to the SAGE method) and multiple corresponding cDNAs comprising at least part of the coding sequence. DADA is a digital method which identifies and counts the abundance of genes by means of restriction fragment fingerprinting.