Genetic and epigenetic variation is a key determinant of human diversity and has an impact on disease predisposition. Single nucleotide polymorphisms (SNPs) and copy number polymorphisms (CNPs) are the main forms of genetic variation. The challenge is to distinguish normal variations from disease-associated changes. Combination of genetic and epigenetic alterations, often together with an environmental component, can cause cancer. In paper I, we investigated possible alterations affecting the transcriptional regulation of PDGFRα in patients affected with central nervous system tumors by characterizing the haplotype combinations in the PDGFRA gene promoter. A specific over-representation of one haplotype (H2δ) in primitive neuroectodermal tumors and ependymomas was observed, suggesting a functional role for the ZNF148/PDGFRα pathway in the tumor pathogenesis. In paper II, 50 glioblastomas were analyzed for DNA copy number variation with a chromosome 22 tiling genomic array. While 20% of tumors displayed monosomy 22, copy number variations affecting a portion of chromosome 22 were found in 14% of cases. This implies the presence of genes involved in glioblastoma development on 22q. Paper III described the analysis of copy number variation of 37 ependymomas using the same array. We detected monosomy in 51.5% of the samples. In addition, we identified two overlapping germline deletions of 2.2 Mb and 320 kb (the latter designated as Ep CNP). In order to investigate whether Ep CNP was a common polymorphism in the normal population or had an association with ependymoma development, we constructed a high-resolution PCR product-based microarray covering this locus (paper IV). For this purpose, we developed a program called Sequence Allocator, which automates the process of array design. This approach allowed assessment of copy number variation within regions of segmental duplications. Our results revealed that gains or deletions were identical in size and encompassed 290 kb. Therefore, papers I-IV suggest that some SNPs and CNPs can be regarded as tumor-associated polymorphisms. Finally, paper V describes variation of DNA methylation among fully differentiated tissues by using an array covering ~9% of the human genome. Major changes in the overall methylation were also found in colorectal cancer cell lines lacking one or two DNA methyltransferases.

Francisella tularensis is the causative agent of the debilitating, sometimes fatal zoonotic disease tularemia. To date, little information has been available on the genetic makeup of this pathogen, its evolution, and the genetic differences which characterize subspecific lineages. These are the main areas addressed in this thesis. The work indicated a high degree of genetic conservation of F. tularensis, both on the sequence level as determined by sequencing and on the compositional level, determined by array-based comparative genomic hybridizations (aCGH). One striking finding was that subsp. mediasiatica was most similar to subsp. tularensis, despite their natural confinement to Central Asia and North America, respectively. All genetic Regions of Difference RD found by aCGH distinguishing lineages were had resulted from repeat-mediated excision of DNA. This was used to identify additional RDs. Such data along with a multiple locus sequence analysis suggested an evolutionary scenario for F. tularensis. Based on genomic information, a novel typing scheme for F. tularensis was furthermore devised and evaluated. This method provided increased robustness compared to previously used methods for F. tularensis typing, while retaining a capacity for high resolution. Finally, the genomic sequence of the highly virulent F. tularensis strain SCHU S4 was determined and analysed. Evidenced by numerous pseudogenes and disrupted metabolic pathways, the bacterium appears to be undergoing a genome reduction process whereby a large proportion of the genetic capacity gradually is lost. It is likely that F. tularensis has irreversibly has evolved into an obligate host-dependent bacterium, incapable of a free-living existence. Unexpectedly, the bacterium was found to be devoid of common virulence mechanisms such as classic toxins, or type III and IV secretion systems. Instead, the virulence of this bacterium is probably largely the result of specific and unusual mechanisms.

Bacteria of the genus Bartonella inhabit the red blood cells of many mammals, including humans, and are transmitted by blood-sucking arthropod vectors. Different species of Bartonella are associated with different mammalian host species, to which they have adapted and normally do not cause any symptoms. Incidental infection of other hosts is however often followed by various disease symptoms, and several Bartonella species are considered as emerging human pathogens. In this work, I have studied the genomic diversity within and between different Bartonella species, with focus on the feline-associated human pathogen B. henselae and its close relatives, the similarly feline-associated B. koehlerae and the trench-fever agent B. quintana which is restricted to humans. In B. henselae, the overall variability in sequence and genome content was modest and well correlated, suggesting low levels of intra-species recombination in the core genome. The variably present genes were located in the prophage and the genomic islands, which are also absent from B. quintana and B. koehlerae, indicating multiple independent excision events. In contrast, diversity of genome structures was immense and probably associated with rearrangements between the repeated genomic islands located around the terminus of replication, possibly to avoid the host’s immune system. In both B. henselae and the mouse-associated species B. grahamii a large portion of the chromosome was manifold amplified in long-time cultures and packaged into phage particles, allowing for different recombination rates for different chromosomal regions. In B. quintana, diversity was studied by sequencing non-coding spacers. The low variability might be due to the recent emergence of this species. Surprisingly, also this species displayed high variability in genome structures, despite its lack of repeated sequences. The results indicate that genome rearrangements and gain or loss of mobile elements are major mechanisms of evolution in Bartonella.

Padlock probes and proximity ligation are two powerful molecular tools for detection of nucleic acids and proteins, respectively. Both methods result in the formation of DNA reporter molecules upon recognition of specific target molecules. These reporter molecules can be designed to include tag sequences that can be analyzed by techniques for nucleic acid analysis. Herein, I present a dual-tag microarray (DTM) platform that is suitable for high-performance analyses of DNA reporter molecule libraries, generated by padlock and proximity probing reactions. The DTM platform was applied for analysis of mRNA transcripts using padlock probes, and of cytokines using proximity ligation. The platform drastically improved specificity of detection, and it allowed precise measurements of proteins and nucleic acids over wide dynamic ranges. The thesis also presents two techniques for multi-probe analyses of biomolecules: the triple-specific proximity ligation assay (3PLA) for protein analyses, and the spliceotyping assay for mRNA analyses. 3PLA allows highly specific measurements of as little as hundreds of target protein molecules by interrogating three target epitopes simultaneously. In spliceotyping the exon composition of individual transcripts are represented as a series of tag sequences in DNA reporter molecules, via a series of target-dependent ligation reactions. Next, the splicing patterns along individual transcripts can be revealed by amplified single molecule detection and step-wise decoding.

In this thesis, methods to study gene and protein expression in cells and tissues were developed and utilized in combination with protein-specific antibodies, with the overall objective to attain greater understanding of protein function. To analyze protein expression in in vitro cultured cell lines, a cell microarray (CMA) was developed, facilitating antibody-based protein profiling of cell lines using immunohistochemistry (IHC). Staining patterns in cell lines were analyzed using image analysis, developed to automatically identify cells and immunohistochemical staining, providing qualitative and quantitative measurements of protein expression. Quantitative IHC data from CMAs stained with nearly 3000 antibodies was used to evaluate the adequacy of using cell lines as models for cancer tissue. We found that cell lines are homogenous with respect to protein expression profiles, and generally more alike each other, than corresponding cancer cells in vivo. However, we found variability between cell lines in regards to the level of retained tumor phenotypic traits, and identified cell lines with a preserved link to corresponding cancer, suggesting that some cell lines are appropriate model systems for specific tumor types. Specific gene expression patterns were analyzed in vitiligo vulgaris and malignant melanoma. Transcriptional profiling of vitiligo melanocytes revealed dysregulation of genes involved in melanin biosynthesis and melanosome function, thus highlighting some mechanisms possibly involved in the pathogenesis of vitiligo. Two new potential markers for infiltrating malignant melanoma, Syntaxin-7 and Discs large homolog 5, were identified using antibody-based protein profiling of melanoma in a tissue microarray format. Both proteins were expressed with high specificity in melanocytic lesions, and loss of Syntaxin-7 expression was associated with more high-grade malignant melanomas. In conclusion, the combination of antibody-based proteomics and microarray technology provided valuable information of expression patterns in cells and tissues, which can be used to better understand associations between protein signatures and disease.

Livet tas för givet av de flesta. Det finns däremot många som ägnar stora delar av sitt liv för att försöka lösa dess mysterier. En del av lösningen ligger i att förstå hur alla molekyler är sammanlänkade i det gigantiska nätverk som definierar den levande organismen. Under det senaste seklet har en hel del forskning utförts för att kartlägga dessa nätverk. Resultatet av dessa mödor kan vi se i de läkemedel som vi har idag och som har utvecklats för att bota eller åtminstone lindra olika sjukdomar och tillstånd. Dessvärre finns det fortfarande många sjukdomar som är obotliga (t.ex. cancer) och mycket arbete krävs för att förstå dem till fullo och kunna designa framgångsrika behandlingar. Arbetet i denna avhandling beskriver en analytisk plattform som kan användas för att effektivisera kartläggningsprocessen; protein-mikroarrayer. Mikroarrayer är ytor som har mikrometerstora (tusendels millimeter) strukturer i ett regelbundet mönster med möjligheten att studera många interaktioner mellan biologiska molekyler samtidigt. Detta medför snabbare och fler analyser - till en lägre kostnad. Protein-mikroarrayer har funnits i ungefär ett decennium och har följt i fotspåren av de framgångsrika DNA-mikroarrayerna. Man bedömer att protein-mikroarrayerna har en minst lika stor potential som DNA mikroarrayerna då det egentligen är mer relevant att studera proteiner, som är de funktionsreglerande molekylerna i en organism. Vi har i detta arbete tillverkat modellytor för stabil inbindning av proteiner, som lämnar dem intakta, funktionella och korrekt orienterade i ett mikroarray format. Därmed har vi adresserat ett stort problem med protein mikroarrays, nämligen att proteiner är känsliga molekyler och har i många fall svårt att överleva tillverkningsprocessen av mikroarrayerna. Vi har även studerat en metod att tillverka mikroarrayer av proteiner bundna till strukturer, som modellerats att efterlikna cellytor. Detta är särkilt viktigt eftersom många (hälften) av dagens (och säkerligen framtidens) läkemedel är riktade mot att påverka denna typ av proteiner och att studera dessa i sin naturliga miljö är därför väldigt relevant. I ett annat projekt har vi använt protein mikroarrayer för att detektera fyra vanliga droger (heroin, amfetamin, ecstasy och kokain). Detektionen baseras på användandet av antikroppar som lossnar från platser på ytan när de kommer i kontakt med ett narkotikum. Detta koncept kan enkelt utvecklas till att detektera mer än bara fyra droger. Vi har även lyckats att parallellt mäta förekomsten av en annan typ av förening på mikroarray ytan, nämligen det explosiva ämnet trinitrotoluen (TNT). Detta visar på en mångsidig plattform för detektionen av i princip vilken typ av farlig eller olaglig substans som helst - och på en yta! Vi föreställer oss därför att möjliga tillämpningsområden finns inom brottsbekämpning, i kampen mot terrorism och mot narkotikamissbruk etc. Mikroarrayerna har i denna avhandling utforskats med optiska metoder som tillåter studie av omärkta proteiner, vilket resulterar i så naturliga molekyler som möjligt. / Life is a thing taken for granted by most. However, it is the life-long quest of many to unravel the mysteries of it. Understanding and characterizing the incomprehensively complex molecular interaction networks within a biological organism, which defines that organism, is a vital prerequisite to understand life itself. Already, there has been a lot of research conducted and a large knowledge has been obtained about these pathways over, especially, the last century. We have seen the fruits of these labors in e.g. the development of medicines which have been able to cure or at least arrest many diseases and conditions. However, many diseases are still incurable (e.g. cancer) and a lot more work is still needed for understanding them fully and designing successful treatments. This work describes a generic analytical tool platform for aiding in more efficient (bio)molecular interaction mapping analyses; protein microarray chips. Microarray chips are surfaces with micrometer sized features with the possibility of studying the interactions of many (thousands to tens of thousands) (bio)molecules in parallel. This allows for a higher throughput of analyses to be performed at a reduced time and cost. Protein microarrays have been around for approximately a decade, following in the footsteps of the, so far, more successfully used DNA microarrays (developed in the 1990s). Microarrays of proteins are more difficult to produce because of the more complex nature of proteins as compared to DNA. In our work we have constructed model surfaces which allow for the stable, highly oriented, and functional immobilization of proteins in an array format. Our capture molecules are based on multivalent units of the chelator nitrilotriacetic acid (NTA), which is able to bind histidine-tagged proteins. Furthermore, we have explored an approach for studying lipid membrane bound systems, e.g. receptor-ligand interactions, in a parallelized, microarray format. The approach relies on the addressable, DNA-mediated adsorption of tagged lipid vesicles. In an analogous work we have used the protein microarray concept for the detection of four common narcotics (heroin, amphetamine, ecstasy, and cocaine). The detection is based on the displacement of loosely bound antibodies from surface array positions upon injection of a specific target analyte, i.e. a narcotic substance. The proof-of-concept chip can easily be expanded to monitor many more narcotic substances. In addition, we have also been able to simultaneously detect the explosive trinitrotoluene (TNT) along with the narcotics, showing that the chip is a versatile platform for the detection of virtually any type of harmful or illegal compound. This type of biosensor system is potentially envisaged to be used in the fight against crime, terrorism, drug abuse etc. Infrared reflection absorption spectroscopy together with ellipsometry has been used to characterize molecular layers used in the fabrication processes of the microarray features. Imaging surface plasmon resonance operating in the ellipsometric mode is subsequently used for functional evaluation of the microarrays using a well-defined receptor-ligand model system. This approach allows simultaneous and continuous monitoring of binding events taking place in multiple regions of interest on the microarray chip. A common characteristic of all the instrumentation used is that there is no requirement for labeling of the biomolecules to be detected, e.g. with fluorescent or radioactive probes. This feature allows for a flexible assay design and the use of more native proteins, without any time-consuming pretreatments.

The sequencing of the human genome has changed life science research in many ways. Novel measurement technologies such as microarray expression analysis, genome-wide SNP typing and mass spectrometry are now producing experimental data of extremely high dimensions. While these techniques provide unprecedented opportunities for exploratory data analysis, the increase in dimensionality also introduces many difficulties. A key problem is to discover the most relevant variables, or features, among the tens of thousands of parallel measurements in a particular experiment. This is referred to as feature selection. For feature selection to be principled, one needs to decide exactly what it means for a feature to be ”relevant”. This thesis considers relevance from a statistical viewpoint, as a measure of statistical dependence on a given target variable. The target variable might be continuous, such as a patient’s blood glucose level, or categorical, such as ”smoker” vs. ”non-smoker”. Several forms of relevance are examined and related to each other to form a coherent theory. Each form of relevance then defines a different feature selection problem. The predictive features are those that allow an accurate predictive model, for example for disease diagnosis. I prove that finding redictive features is a tractable problem, in that consistent estimates can be computed in polynomial time. This is a substantial improvement upon current theory. However, I also demonstrate that selecting features to optimize prediction accuracy does not control feature error rates. This is a severe drawback in life science, where the selected features per se are important, for example as candidate drug targets. To address this problem, I propose a statistical method which to my knowledge is the first to achieve error control. Moreover, I show that in high dimensions, feature sets can be impossible to replicate in independent experiments even with controlled error rates. This finding may explain the lack of agreement among genome-wide association studies and molecular signatures of disease. The most predictive features may not always be the most relevant ones from a biological perspective, since the predictive power of a given feature may depend on measurement noise rather than biological properties. I therefore consider a wider definition of relevance that avoids this problem. The resulting feature selection problem is shown to be asymptotically intractable in the general case; however, I derive a set of simplifying assumptions which admit an intuitive, consistent polynomial-time algorithm. Moreover, I present a method that controls error rates also for this problem. This algorithm is evaluated on microarray data from case studies in diabetes and cancer. In some cases however, I find that these statistical relevance concepts are insufficient to prioritize among candidate features in a biologically reasonable manner. Therefore, effective feature selection for life science requires both a careful definition of relevance and a principled integration of existing biological knowledge. / Sekvenseringen av det mänskliga genomet i början på 2000-talet tillsammans och de senare sekvenseringsprojekten för olika modellorganismer har möjliggjort revolutionerade nya biologiska mätmetoder som omfattar hela genom. Microarrayer, mass-spektrometri och SNP-typning är exempel på sådana mätmetoder. Dessa metoder genererar mycket högdimensionell data. Ett centralt problem i modern biologisk forskning är således att identifiera de relevanta variablerna bland dessa tusentals mätningar. Detta kallas f¨or variabelsökning. För att kunna studera variabelsökning på ett systematiskt sätt är en exakt definition av begreppet ”relevans” nödvändig. I denna avhandling behandlas relevans ur statistisk synvinkel: ”relevans” innebär ett statistiskt beroende av en målvariabel ; denna kan vara kontinuerlig, till exempel en blodtrycksmätning på en patient, eller diskret, till exempel en indikatorvariabel såsom ”rökare” eller ”icke-rökare”. Olika former av relevans behandlas och en sammanhängande teori presenteras. Varje relevansdefinition ger därefter upphov till ett specifikt variabelsökningsproblem. Prediktiva variabler är sådana som kan användas för att konstruera prediktionsmodeller. Detta är viktigt exempelvis i kliniska diagnossystem. Här bevisas att en konsistent skattning av sådana variabler kan beräknas i polynomisk tid, så att variabelssökning är möjlig inom rimlig beräkningstid. Detta är ett genombrott jämfört med tidigare forskning. Dock visas även att metoder för att optimera prediktionsmodeller ofta ger höga andelar irrelevanta varibler, vilket är mycket problematiskt inom biologisk forskning. Därför presenteras också en ny variabelsökningsmetod med vilken de funna variablernas relevans är statistiskt säkerställd. I detta sammanhang visas också att variabelsökningsmetoder inte är reproducerbara i vanlig bemärkelse i höga dimensioner, även då relevans är statistiskt säkerställd. Detta förklarar till viss del varför genetiska associationsstudier som behandlar hela genom hittills har varit svåra att reproducera. Här behandlas också fallet där alla relevanta variabler eftersöks. Detta problem bevisas kräva exponentiell beräkningstid i det allmänna fallet. Dock presenteras en metod som löser problemet i polynomisk tid under vissa statistiska antaganden, vilka kan anses rimliga för biologisk data. Också här tas problemet med falska positiver i beaktande, och en statistisk metod presenteras som säkerställer relevans. Denna metod tillämpas på fallstudier i typ 2-diabetes och cancer. I vissa fall är dock mängden relevanta variabler mycket stor. Statistisk behandling av en enskild datatyp är då otillräcklig. I sådana situationer är det viktigt att nyttja olika datakällor samt existerande biologisk kunskap för att för att sortera fram de viktigaste fynden.

One of the most important parameters affecting hydrogen production in photofermentation process is the type of carbon and nitrogen sources. For this reason in this research, the effect of different nitrogen sources (5mM ammonium chloride and 2mM glutamate) and acetate concentrations (40

This thesis analyzes major microarray repositories and presents a metadata framework both to address the current issues and to promote the main operations such as knowledge discovery, sharing, integration, and exchange. The proposed framework is demonstrated in a case study on real data and can be used for other high throughput repositories in biomedical domain. Not only the number of microarray experimentation increases, but also the size and complexity of the results rise in response to biomedical inquiries. And, experiment results are significant when examined in a batch and placed in a biological context. There have been standardization initiatives on content, object model, exchange format, and ontology. However, they have proprietary information space. There are backlogs and the data cannot be exchanged among the repositories. There is a need for a format and data management standard at present.iv v We introduced a metadata framework to include metadata card and semantic nets to make the experiment results visible, understandable and usable. They are encoded in standard syntax encoding schemes and represented in XML/RDF. They can be integrated with other metadata cards, semantic nets and can be queried. They can be exchanged and shared. We demonstrated the performance and potential benefits with a case study on a microarray repository. This study does not replace any product on repositories. A metadata framework is required to manage such huge data. We state that the backlogs can be reduced, complex knowledge discovery queries and exchange of information can become possible with this metadata framework.

Francisella tularensis is a gram-negative bacterium that is the cause of the serious and sometimes fatal disease, tularemia, in a wide range of animal species and in humans. The response of cells of the mouse macrophage cell line J774 to infection with Francisella tularensis LVS was analyzed by means of a DNA microarray. It was observed that the infection conferred an oxidative stress upon the target cells and many of the host defense mechanisms appeared to be intended to counteract this stress. The infection was characterized by a very modest inflammatory response. Tularemia caused by inhalation of F. tularensis subspecies tularensis is one of the most aggressive infectious diseases known. We used the mouse model to examine in detail the host immune response in the lung. After an aerosol challenge all mice developed clinical signs of severe disease, showed weight loss by day four of infection, and died the next day. Gene transcriptional changes in the mouse lung samples were examined on day one, two, and four of infection. Genes preferentially involved in host immune responses were activated extensively on day four but on day one and two, only marginally or not at all. Several genes upregulated on day four are known to depend on IFN-gamma or TNF-alpha for their regulation. In keeping with this finding, TNF-alpha and IFN-gamma levels were found to be increased significantly in bronchoalveolar lavage on day four. We undertook an analysis of the transcriptional response in peripheral blood during the course of ulceroglandular tularemia by use of Affymetrix microarrays. Samples were obtained from seven individuals at five occasions during two weeks after the first hospital visit and convalescent samples three months later. In total 265 genes were differentially expressed. The most prominent changes were noted in samples drawn on days 2-3 and a considerable proportion of the upregulated genes appeared to represent an IFN-gamma-induced response and also a pro-apoptotic response. Genes involved in the generation of innate and acquired immune responses were found to be downregulated, presumably a pathogen-induced event. A logistic regression analysis revealed that seven genes were good predictors of the early phase of tularemia. Recently, a large number of methods for the analysis of microarray data have been proposed but there are few comparisons of their relative performances. We undertook a study to evaluate established and novel methods for filtration, background adjustment, scanning, and censoring. For all analyses, the sensitivities at low false positive rates were observed together with a bias measurement. In general, there was a trade off between the analyses ability to identify differentially expressed genes and their ability to obtain unbiased estimators of the desired ratios. A commonly used standard analysis using background adjustment performed poorly. Interestingly, the constrained model combining data from several scans resulted in high sensitivities. For experiments where only low false discovery rates are acceptable, the use of the constrained model or the novel partial filtration method are likely to perform better than some commonly used standard analyses.