Optimization of Oligonucleotide Sets for DNA Microarrays

DNA microarrays or DNA chips are a rapidly developing technology in the domain of nucleic acid analytics. They facilitate rapid, parallel, cheap and highly specific and sensitive assays for the analysis of biodiversity or genetic profiles. Thousands of reactions with short nucleic acid molecules (so called oligonucleotides or oligos) have to be optimized, and parameters and criteria have to be considered. This chip design or configuration, is the subject of this work.The oligos are designed to detect the presence of a given set of target sequences. Provided that an oligo gives a signal by hybridization and the target sequence is in the sample, this signal is called true positive. The detection of non target sequences, e.g. human DNA within a virus assay, is called false positive signal or cross hybridization. A statistic on these types of signals determine the specificity and sensitivity of the oligos or a whole oligo set. Another criterion is the hybridization efficiency, which is influenced by a lot of properties of the oligo and the target sequence, like secondary structures. To detect and cover all variations of highly variable virus genomes it is necessary to design a combinatorily optimized oligo set. This work shows three approaches for this set cover problem , considering the above mentioned criteria and properties of the oligos or the whole set. These are a modified greedy search, a combination of gradient descent and competition and genetic algorithms.A web based bioinformatics system optiNA optimal Nucleic Acids was developed that enables the design of combinatorily optimized oligo sets. It reduces the development process to a few days, there is the possibility to process large amounts of sequence data and work with a lot of parameters and conditions and elements of the hybridization protocol can be used to parametrize optiNA. The error-prone manual work with large amounts of data is unnecessary