Motivation: As the human genome project progresses and some microbial and eukaryotic genomes are recognized, numerous biotechnological processes have attracted increasing number of biologists, bioengineers and computer scientists recently. Biotechnological processes profoundly involve production and analysis of highthroughput experimental data. Numerous sequence libraries of DNA and protein structures of a large number of micro-organisms and a variety of other databases related to biology and chemistry are available. For example, microarray technology, a novel biotechnology, promises to monitor the whole genome at once, so that researchers can study the whole genome on the global level and have a better picture of the expressions among millions of genes simultaneously. Today, it is widely used in many fields- disease diagnosis, gene classification, gene regulatory network, and drug discovery. For example, designing organism specific microarray and analysis of experimental data require combining heterogeneous computational tools that usually differ in the data format; such as, GeneMark for ORF extraction, Promide for DNA probe selection, Chip for probe placement on microarray chip, BLAST to compare sequences, MEGA for phylogenetic analysis, and ClustalX for multiple alignments. Solution: Surprisingly enough, despite huge research efforts invested in DNA array applications, very few works are devoted to computer-aided optimization of DNA array design and manufacturing. Current design practices are dominated by ad-hoc heuristics incorporated in proprietary tools with unknown suboptimality. This will soon become a bottleneck for the new generation of high-density arrays, such as the ones currently being designed at Perlegen [109]. The goal of the already accomplished research was to develop highly scalable tools, with predictable runtime and quality, for cost-effective, computer-aided design and manufacturing of DNA probe arrays. We illustrate the utility of our approach by taking a concrete example of combining the design tools of microarray technology for Harpes B virus DNA data.

Developments within the field of genetic analysis have during the last decade become enormous. Advances in DNA sequencing technology have increased throughput from a thousand bases to over a billion bases in a day and decreased the cost thousandfold per base. Nevertheless, to sequence complex genomes like the human is still very expensive and efforts to attain even higher throughputs for less money are undertaken by researchers and companies. Genotyping systems for single nucleotide polymorphism (SNP) analysis with whole genome coverage have also been developed, with low cost per SNP. There is, however, a need for genotyping assays that are more cost efficient per sample with considerably higher accuracy. This thesis is focusing on a technology, based on competitive allele-specific extension and microarray detection, for genetic analysis. To increase specificity in allele-specific extension (ASE), a nucleotide degrading enzyme, apyrase, was introduced to compete with the polymerase, only allowing the fast, perfect matched primer extension to occur. The aim was to develop a method for analysis of around twenty loci in hundreds of samples in a high-throughput microarray format. A genotyping method for human papillomavirus has been developed, based on a combination of multiplex competitive hybridization (MUCH) and apyrase-mediated allele-specific extension (AMASE). Human papillomavirus (HPV), which is the causative agent in cervical cancer, exists in over a hundred different types. These types need to be determined in clinical samples. The developed assay can detect the twenty-three most common high risk types, as well as semi-quantifying multiple infections, which was demonstrated by analysis of ninety-two HPV-positive clinical samples. More stringent conditions can be obtained by increased reaction temperature. To further improve the genotyping assay, a thermostable enzyme, protease, was introduced into the allele-specific extension reaction, denoted PrASE. Increased sensitivity was achieved with an automated magnetic system that facilitates washing. The PrASE genotyping of thirteen SNPs yielded higher conversion rates, as well as more robust genotype scoring, compared to ASE. Furthermore, a comparison with pyrosequencing, where 99.8 % of the 4,420 analyzed genotypes were in concordance, indicates high accuracy and robustness of the PrASE technology. Single cells have also been analyzed by the PrASE assay to investigate loss of alleles during skin differentiation. Single cell analysis is very demanding due to the limited amounts of DNA. The multiplex PCR and the PrASE assay were optimized for single cell analysis. Twenty-four SNPs were genotyped and an increased loss of genetic material was seen in cells from the more differentiated suprabasal layers compared to the basal layer. / QC 20100714