Hand osteoarthritis (OA) is a degenerative joint disease that causes pain, functional limitation and negatively affects the patients' quality of life. The most severe subtype of this disease is erosive OA. Erosive hand OA is characterized by an abrupt onset, inflammation and is linked to worse outcomes than non-erosive hand OA. Current methods do not allow early diagnosis or to distinguish between patients with different forms at disease onset. This could be changed by the utilization of biomarkers in clinical practice. Biomarkers are molecules released into circulation that reflect biological processes. The main goal of this study was to analyze the levels of circulating biomarkers with the aim to differentiate patients from healthy subjects and patients with erosive OA from patients with non-erosive disease. Serum concentrations of seven biomarkers and the expression of plasma microRNAs were determined. Patients with hand OA showed altered cartilage metabolism, increased levels of adiponectin, decreased levels of clusterin and a dysregulated expression of several microRNAs in comparison to the healthy population. Patients with erosive OA had lower levels of clusterin and decreased expression of miR-151-3p than those with the non-erosive form of the disease. These findings suggest the potential...

In this dissertation we present rule-based machine learning methods for solving problems with high-dimensional or complex datasets. We are applying decision tree methods on blood-based biomarkers and neuropsychological tests to predict Alzheimer’s disease in its early stages. We are also using tree-based methods to identify disparity in dementia related biomarkers among three female ethnic groups. In another part of this research, we tried to use rule-based methods to identify homogeneous subgroups of subjects who share the same risk patterns out of a heterogeneous population. Finally, we applied a network-based method to reduce the dimensionality of a clinical dataset, while capturing the interaction among variables. The results show that the proposed methods are efficient and easy to use in comparison to the current machine learning methods.

No description available.

The aim of this thesis was to investigate the associations between fatty acid (FA) biomarkers, carbohydrate intake, and abdominal obesity (AO) and related anthropometric measures in a population-based cohort of men and women in Stockholm County. The overall hypothesis was that dietary fat quality assessed by serum and adipose tissue FA composition, and dietary intake of especially carbohydrates is associated with AO. FA composition was assessed by liquid gas chromatography, and AO was measured as waist circumference (WC), waist hip ratio (WHR) and sagittal abdominal diameter (SAD). Dietary intake was assessed by 7-day food records. Papers I, II, III, and IV were all observational studies based on a Swedish population in Stockholm County (n=5460). A sub-cohort of only men (n=301) was included in Papers II, III, and IV. In Paper I, serum proportions of the polyunsaturated FA (PUFA), linoleic acid (LA) (18:2n6), was inversely associated with AO in both men and women, whereas a positive association was observed between the saturated FA (SFA), palmitic acid (PA) (16:0) and AO measures. These findings support recent interventional studies suggesting that a higher relative intake of PUFA (LA) from vegetable oils as compared with 16:0 is associated with decreased abdominal adiposity. In Paper II, we investigated whether biomarkers of dietary fat quality were related to the corresponding FA intake from fat-rich foods reported in a short food frequency questionnaire (FFQ). Serum proportions of the long-chain n-3 PUFAs eicosapentaenoic acid (EPA) and docosahexanoic acid (DHA) were higher among men with higher total fish intake. Serum LA was higher among men who reported a consumption of more than 5 g/d of margarine. Absolute agreement between intakes assessed with FFQ of 60YO and 7-day food record of "Kost och Metabola syndromet"/"Diet and the Metabolic syndrome" (KOMET) was highest for alcohol, total fish, and eggs. Weighted Kappa statistics revealed the strongest agreement for alcohol, margarine, and fruits. In Paper III, carbohydrate intake was inversely associated with 16:0 in serum phospholipids (PL). Disaccharide and alcohol intake was positively and non-linearly associated with palmitoleic acid (16:1) and stearoyl-CoA-desaturase (SCD) activity in PL. Alcohol was consistently associated with higher SFA and monounsaturated FA (MUFA). Results of Paper IV indicated that total carbohydrate intake was inversely associated with measures of AO and central fat distribution, WHR and SAD, respectively. Likewise, monosaccharide intake was associated with lower AO. In contrast, alcohol intake was associated with AO prevalence and all anthropometric measurements. In conclusion, serum SFA (palmitic acid) was positively associated with AO, whereas n-6 PUFA (linoleic acid) was associated with lower AO. High intake of total carbohydrate and monosaccharides were associated with lower AO. Overall, these results support a beneficial role on adiposity of diets that are higher in polyunsaturated fat (vegetable oils) and total carbohydrates compared with saturated fat.

The long history of co-existence of bacteria and their protozoan predators in aquatic environments has led to evolution of protozoa resistant bacteria (PRB). Many of these bacteria are also pathogenic to humans. However, the ecological drivers determining the occurrence of different types of PRB in aquatic environments, and the eco-evolutionary link between bacterial adaptation and the resulting implications for mammalian hosts are poorly known. This thesis examines the impact of nutrients and predation on PRB, as well as the ecological and evolutionary connection between their life in aquatic environments and mammalian hosts. In the first study seven bacterial isolates from the Baltic Sea were investigated for their plasticity of adaptation to predation. The response to predation showed large variation where some bacteria rapidly developed a degree of grazing resistance when exposed to predators. The rapid adaptation observed may result in bacterial communities being resilient or resistant to predation, and thus rapid adaptation may be a structuring force in the food web. With the aim to elucidate the link between occurrence of PRB and environmental conditions, a field study and a laboratory experiment were performed. In both studies three PRB genera were found: Mycobacterium, Pseudomonas and Rickettsia. PRB were found both in oligotrophic and eutrophic waters, indicating that waters of all nutrient states can harbor pathogenic bacteria. However, the ecological strategy of the PRB varied depending on environmental nutrient level and disturbance. Using an advanced bioinformatic analysis, it was shown that ecotypes within the same PRB genus can be linked to specific environmental conditions or the presence of specific protozoa, cyanobacteria or phytoplankton taxa. These environmental conditions or specific plankton taxa could potentially act as indicators for occurrence of PRB. Finally, using four mutants (with specific protein deletions) of the pathogenic and predation resistant Francisella tularensis ssp. holarctica, I found evidence of an eco-evolutionary connection between the bacterium´s life in aquatic and mammalian hosts (aquatic amoeba Acanthamoeba castellanii and a murine macrophage).  To a large extent F. t. holarctica use similar mechanisms to persist predation by protozoa and to resist degradation by mammal macrophages. To summarize I found a link between predation resistant bacteria in aquatic environments and bacteria that are pathogenic to mammals. Further, I showed that different environmental conditions rapidly selects for PRB with either intracellular or extracellular lifestyles. This thesis provides insights regarding environmental conditions and biomarkers that can be used for assessment of aquatic environments at risk for spreading pathogenic bacteria. / <p>Medfinansiärer var även: Swedish Ministry of Defence (A4040, A4042, A404215, A404217), Swedish Minestry of Foreign Affairs (A4952), Swedish Civil Contingencies Agency (B4055)</p>

>Magister Scientiae - MSc / Type 2 diabetes mellitus (T2DM), a metabolic disease characterized by chronic hyperglycemia, is the most prevalent form of diabetes globally, affecting approximately 95 % of the total number of people with diabetes i.e. approximately 366 million. Furthermore, it is also the most prevalent form in South Africa (SA), affecting approximately 3.5 million individuals. This disease and its adverse complications can be delayed or prevented if detected early. Standardized diagnostic tests for T2DM have a few limitations which include the inability to predict the future risk of normal glucose tolerance individuals developing T2DM, they are dependent on blood glucose concentration, its invasiveness, and they cannot specify between T1DM and T2DM. Therefore, there is a need for biomarkers which could be used as a tool for the early and specific detection of T2DM. MicroRNAs are small non-coding RNA molecules which play a key role in controlling gene expression and certain biological processes. Studies show that dysregulation of microRNAs may lead to various diseases including T2DM, and thus, may be useful biomarkers for disease detection. Therefore, identifying biomarkers like microRNAs as a tool for the early and specific detection of T2DM, have great potential for diagnostic purposes. The main focus of this investigation, therefore, is the early detection of T2DM by the identification and validation of novel biomarkers. Furthermore, based on previous studies, the aim of the investigation was to identify differentially expressed miRNAs as well as identify their potential target genes associated with the onset and progression of T2DM. An in silico approach was used to identify miRNAs found to be differentially expressed in the serum/plasma of T2DM individuals. Three publically available target prediction software were used for target gene prediction of the identified miRNA. The target genes were subjected to functional analysis using a web-based software, namely DAVID. Functions which were clustered with an enrichment score > 1.3 were considered significant. The ranked target genes mostly had gene ontologies linked with “transcription regulation”, “neuron signalling, and “metal ion binding”. The ranked target genes were then split into two lists – an up-regulated (ur) miRNA targeted gene list and a down-regulated (dr) miRNA targeted gene list. The in silico method used in this investigation produced a final total of 4 miRNAs: miR-dr-1, miR-ur-1, miR-ur-2, and miR-ur-3. Based on the bioinformatics results, miR-dr-1 and its target genes LDLR, PPARA and CAMTA1, seemed the most promising miRNA for biomarker validation, due to the function of the target genes being associated with T2DM onset and progression. The expression levels of the miRNAs were then profiled in kidney tissue of male Wistar rats that were on a high fat diet (HFD), streptozotocin (STZ)-induced T1DM, and non-diabetic control rats via qRT-PCR analysis. The hypothesis was that similar miRNA expression would be found in the HFD kidney samples compared to serum expression levels of the miRNA obtained from the two databases, since kidneys are involved in cleansing the blood from impurities. This hypothesis proved to be true for all miRNAs except for miR-ur-2. Additionally, miR-ur-1 seemed the most significant miRNA due to it having different expression ratios for T1DM and T2DM (i.e. -7.65 and 4.2 fold, respectively). Future work, therefore, include validation of the predicted target genes to the miRNAs of interest i.e. miR-dr-1: PPARA and LDLR and miR-ur-1: CACNB2, using molecular approaches such as the luciferase assays and western blots.

Introduction: Atherosclerosis remains asymptomatic until it progresses to cause flow-limiting disease. Identifying patients at high risk in the early stages of the atherosclerotic process may allow modification of cardiovascular risk by effective preventive strategies. Various non-invasive tests have been studied and have shown promising results in predicting future adverse cardiovascular events. The objective of this study was to establish the carotid ultrasonographic markers that best correlate with angiographic coronary artery disease (CAD) and the relationship between left ventricular geometry, carotid atherosclerosis, biomarkers and CAD in patients with unobstructed coronary arteries, chronic stable angina (CSA) and acute coronary syndromes (ACS). Methods: Carotid ultrasound examination, echocardiography and serum biomarker estimation were performed in consecutive patients who underwent coronary angiography for evaluation of stable or acute chest pain. Results: A total of 146 subjects were recruited into the study with a mean age of 56.9 ± 10.6 (range 29 to 85) years; 120 were men (82%) and 26 (18%) women. Twenty-one percent of the study population had unobstruced coronaries, 42% had stable CAD and 37% had presented with ACS. There was no significant difference in the carotid intima media thickness (CIMT) measurements between the three groups. CIMT correlated with abnormal left ventricular geometry but not with the presence or severity of CAD. The presence of carotid plaque and plaque score correlated with obstructive CAD, but was not significantly different between stable CAD and ACS patients. There was a trend towards more echogenic plaque in the stable CAD group. The composite score of IMT and plaque was positively correlated with the presence and severity of CAD. The averaged myocardial peak systolic and early diastolic velocities were significantly lower in those with obstructive CAD. CRP and osteopontin levels were higher in the ACS patients. Conclusions: Carotid plaque and not CIMT was associated with angiographic coronary artery disease. Averaged systolic and early diastolic myocardial velocities by tissue doppler imaging correlated with obstructive CAD. Novel serum biomarkers are promising and further studies are needed.

La sclérose en plaques (SEP) est une maladie inflammatoire auto-immune du SNC qui affecte 90 000 personnes en France et génère un coût important pour la société. La forme la plus fréquente (85% des cas) est la SEP récurrente-rémittente (RRMS) caractérisée par des poussées démyélinisantes entrecoupées de périodes de rémission après un syndrome cliniquement isolé (CIS). La conversion en RRMS après un CIS, caractérisée par une nouvelle poussée ou de nouvelles lésions sur les IRM de suivi, survient après un délai variant de quelques mois à plus de 10 ans (20% des cas, « SEP bénigne »). Les traitements disponibles (immunomodulateurs et immunosuppresseurs) ont une efficacité indéniable pour la prévention des poussées et doivent être initiés dès le diagnostic pour une efficacité maximale, surtout chez les patients montrant une conversion rapide en SEP après un CIS. Leur intérêt est en revanche négligeable pour les formes bénignes. Ainsi, il existe un besoin d’identifier des biomarqueurs pronostiques précoces de la SEP pour une prise en charge optimale des patients. Le but de ma thèse est d’identifier des biomarqueurs pronostiques de SEP en utilisant différentes approches de protéomique quantitative,. Dans un premier temps, nous avons utilisé des échantillons cliniques de LCR de patients. Ces travaux m’ont permis d’identifier, en protéomique quantitative utilisant un marquage chimique des peptides (TMT), plusieurs candidats biomarqueurs de SEP. Ceux-ci incluent deux chitinase-3-like protéines, CHI3L1 et CHI3L2 dont le taux mesuré en ELISA, dans le LCR, augmente avec l’évolution de la maladie. Ces travaux ont également démontré que la concentration sanguine de CHI3L1 constituait un biomarqueur de la progression de la maladie. Dans un second temps, nous avons combiné l’analyse du LCR de différents patients SEP et contrôles en protéomique Label-Free avec un modèle préclinique analysant les modifications du sécrétome d’oligodendrocytes murins en culture primaire par approche SILAC. Les biomarqueurs potentiels, 87 protéines correspondant à 226 peptides cibles, ont été ensuite vérifiés en protéomique quantitative ciblée ou parallel reaction monitoring (PRM), à l’aide de peptides marqués servant de référence. Les 11 candidats les plus discriminants issus de cette étape ont ensuite été vérifiés en PRM sur une cohorte plus large d’échantillons de patients. Finalement, 7 candidats ont montré un profil significatif au cours de cette première validation. Ces candidats biomarqueurs permettent notamment de discriminer les différentes formes évolutives de la SEP et de distinguer cette pathologie d’autres maladies inflammatoires et neurologiques. De plus, ces derniers pourraient avoir un intérêt prognostique permettant d’identifier les patients qui vont développer une SEP après la découverte fortuite de lésions typiques de la maladie à l’IRM. Ce travail de thèse a donc caractérisé de nouveaux biomarqueurs de SEP devant être validés sur de larges cohortes multicentriques de patients et ouvre de nouvelles perspectives sur la compréhension des mécanismes physiopathologiques de la SEP. / Discovery, confirmation and verification of candidate biomarkers for multiple sclerosis diagnosis and prognosis in cerebrospinal fluid.Multiple sclerosis (MS) is an inflammatory disease of the central nervous system. Most often the disease initiates by a first demyelinating event called clinically isolated syndrome (CIS), followed by remission periods and relapses occurring at irregular intervals. Clinical symptoms and MRI are used for diagnosis, but clinicians lack tools to predict the rate of disease progression. This study aims at identifying biomarkers that predict the delay of conversion to MS after a CIS. We compared the cerebrospinal fluid (CSF) proteome from MS patients and symptomatic controls and the CSF from CIS patients with rapid conversion to MS (<1 year) and CIS patients with slow conversion to MS (>2 years). For the discovery step, human CSF samples (n=40) depleted of the 20 major plasma proteins were digested using a modified filter-assisted sample preparation (FASP) and analysed by high-resolution mass spectrometry using isobaric mass tag labelling or label-free quantification procedures. Proteins upregulated in CSF from MS patients included two proteins involved in tissue remodeling, namely chitinase-3-like protein-1 (CHI3L1) and chitinase-3-like protein-2 (CHI3L2). Their increased level in CSF of MS patients was confirmed by ELISA in a new cohort comprising CIS and MS patients (n=123) at different disease stages. Moreover, CHI3L1 levels in CSF and serum from CIS patients discriminated patients with rapid conversion to MS (< one year) from those with slower conversion.We also implemented a PRM method (peptide selection, dilution optimization of heavy isotope labeled non-purified peptides, reproducibility evaluation and method validation) to qualify a larger set of candidate biomarkers (226 peptides corresponding to 87 proteins) in a cohort different from the one used for the discovery step (n=60), including CSF from controls and MS patients at different disease stages. Finally, to further verify the 11 candidate biomarkers that passed this qualification step, we monitored 16 peptides in a new PRM assay, using shorter gradient and high-purity heavy isotope labeled peptides. This new PRM analysis was performed on a larger cohort (n=189) that included CSF of patients with other inflammatory and non-inflammatory neurological disorders in addition to control and MS patients. These analyses identified seven robust candidate biomarkers, which might help to discriminate patients suffering from MS or other neurological disorders.

In this dissertation, we discuss a reading profiling system, a biological data visualization system and a sports visualization system. Self-tracking is getting increasingly popular in the field of personal informatics. Reading profiling can be used as a personal data collection method. We present UUAT, an unintrusive user attention tracking system. In UUAT, we used user interaction data to develop technologies that help to pinpoint a users reading region (RR). Based on computed RR and user interaction data, UUAT can identify a readers reading struggle or interest. A biomarker is a measurable substance that may be used as an indicator of a particular disease. We developed CancerVis for visual and interactive analysis of cancer data and demonstrate how to apply this platform in cancer biomarker research. CancerVis provides interactive multiple views from different perspectives of a dataset. The views are synchronized so that users can easily link them to a same data entry. Furthermore, CancerVis supports data mining practice in cancer biomarker, such as visualization of optimal cutpoints and cutthrough exploration. Tennis match summarization helps after-live sports consumers assimilate an interested match. We developed TennisVis, a comprehensive match summarization and visualization platform. TennisVis offers chart- graph for a client to quickly get match facts. Meanwhile, TennisVis offers various queries of tennis points to satisfy diversified client preferences (such as volley shot, many-shot rally) of tennis fans. Furthermore, TennisVis offers video clips for every single tennis point and a recommendation rating is computed for each tennis play. A case study shows that TennisVis identifies more than 75% tennis points in full time match.

The human genome contains tens of thousands of gene loci which code for an even greater number of protein and RNA products. The highly complex temporal and spatial expression of these genes makes possible all the biological processes of life. Altered gene expression by mutation or deregulation is fundamental for the development of many human diseases. The ultimate aim of this thesis was to identify gene expression changes relevant to cancer. The advent of genome-wide expression profiling techniques, such as microarrays, has provided powerful new tools to identify such changes and researchers are now faced with an explosion of gene expression data. Processing, comparing and integrating these data present major challenges. I approached these challenges by developing and assessing novel methods for cross-platform analysis of expression data, scalable subspace clustering, and curation of experimental gene regulation data from the published literature. I found that combining results from different expression platforms increases reliability of coexpression predictions. However, I also observed that global correlation between platforms was generally low, and few gene pairs reached reasonable thresholds for high-confidence coexpression. Therefore, I developed a novel subspace clustering algorithm, able to identify coexpressed genes in experimental subsets of very large gene expression datasets. Biological assessment against several metrics indicates that this algorithm performs well. I also developed a novel meta-analysis method to identify consistently reported genes from differential expression studies when raw data are unavailable. This method was applied to thyroid cancer, producing a ranked list of significantly over-represented genes. Tissue microarray analysis of some of these candidates and others identified a number of promising biomarkers for diagnostic and prognostic classification of thyroid cancer. Finally, I present ORegAnno (www.oreganno.org), a resource for the community-driven curation of experimentally verified regulatory sequences. This resource has proven a great success with ~30,000 sequences entered from over 900 publications by ~50 contributing users. These data, methods and resources contribute to our overall understanding of gene regulation, gene expression, and the changes that occur in cancer. Such an understanding should help identify new cancer mechanisms, potential treatment targets, and have significant diagnostic and prognostic implications. / Medicine, Faculty of / Medical Genetics, Department of / Graduate