In this thesis, we address the identification of biomarkers in high-dimensional omics data. The identification of valid biomarkers is especially relevant for personalized medicine that depends on accurate prediction rules. Moreover, biomarkers elucidate the provenance of disease, or molecular changes related to disease. From a statistical point of view the identification of biomarkers is best cast as variable selection. In particular, we refer to variables as the molecular attributes under investigation, e.g. genes, genetic variation, or metabolites; and we refer to observations as the specific samples whose attributes we investigate, e.g. patients and controls. Variable selection in high-dimensional omics data is a complicated challenge due to the characteristic structure of omics data. For one, omics data is high-dimensional, comprising cellular information in unprecedented details. Moreover, there is an intricate correlation structure among the variables due to e.g internal cellular regulation, or external, latent factors. Variable selection for uncorrelated data is well established. In contrast, there is no consensus on how to approach variable selection under correlation. Here, we introduce a multivariate framework for variable selection that explicitly accounts for the correlation among markers. In particular, we present two novel quantities for variable importance: the correlation-adjusted t (CAT) score for classification, and the correlation-adjusted (marginal) correlation (CAR) score for regression. The CAT score is defined as the Mahalanobis-decorrelated t-score vector, and the CAR score as the Mahalanobis-decorrelated correlation between the predictor variables and the outcome. We derive the CAT and CAR score from a predictive point of view in linear discriminant analysis and regression; both quantities assess the weight of a decorrelated and standardized variable on the prediction rule. Furthermore, we discuss properties of both scores and relations to established quantities. Above all, the CAT score decomposes Hotelling’s T 2 and the CAR score the proportion of variance explained. Notably, the decomposition of total variance into explained and unexplained variance in the linear model can be rewritten in terms of CAR scores. To render our approach applicable on high-dimensional omics data we devise an efficient algorithm for shrinkage estimates of the CAT and CAR score. Subsequently, we conduct extensive simulation studies to investigate the performance of our novel approaches in ranking and prediction under correlation. Here, CAT and CAR scores consistently improve over marginal approaches in terms of more true positives selected and a lower model error. Finally, we illustrate the application of CAT and CAR score on real omics data. In particular, we analyze genomics, transcriptomics, and metabolomics data. We ascertain that CAT and CAR score are competitive or outperform state of the art techniques in terms of true positives detected and prediction error.

Although human biomedical and physiological information is readily available, such information for great apes is limited. We analyzed clinical chemical biomarkers in serum samples from 277 wild- and captive-born great apes and from 312 healthy human volunteers as well as from 20 rhesus macaques. For each individual, we determined a maximum of 33 markers of heart, liver, kidney, thyroid and pancreas function, hemoglobin and lipid metabolism and one marker of inflammation. We identified biomarkers that show differences between humans and the great apes in their average level or activity. Using the rhesus macaques as an outgroup, we identified human-specific differences in the levels of bilirubin, cholinesterase and lactate dehydrogenase, and bonobo-specific differences in the level of apolipoprotein A-I. For the remaining twenty-nine biomarkers there was no evidence for lineage-specific differences. In fact, we find that many biomarkers show differences between individuals of the same species in different environments. Of the four lineagespecific biomarkers, only bilirubin showed no differences between wild- and captive-born great apes. We show that the major factor explaining the human-specific difference in bilirubin levels may be genetic. There are human-specific changes in the sequence of the promoter and the protein-coding sequence of uridine diphosphoglucuronosyltransferase 1 (UGT1A1), the enzyme that transforms bilirubin and toxic plant compounds into water-soluble, excretable metabolites. Experimental evidence that UGT1A1 is down-regulated in the human liver suggests that changes in the promoter may be responsible for the human-specific increase in bilirubin. We speculate that since cooking reduces toxic plant compounds, consumption of cooked foods, which is specific to humans, may have resulted in relaxed constraint on UGT1A1 which has in turn led to higher serum levels of bilirubin in humans.

Le long de la frontière entre la Thaïlande et le Myanmar (TMB), le paludisme se caractérise par une forte hétérogénéité de la transmission, une forte prévalence en porteurs sub-microscopiques et par l’émergence de la résistance à l’artémisinine chez Plasmodium falciparum. L'identification précoce des « foyers » infectieux et leurs éliminations sont nécessaires pour contenir la résistance à l'artémisinine. L'objectif de cette thèse était de démontrer l’intérêt d’utiliser des biomarqueurs sérologiques de l'exposition humaine aux piqûres d'anophèles (gSG6-P1) et au Plasmodium (CSP & MSP1-19) pour quantifier le contact homme-vecteur et identifier les foyers résiduels de transmission. Des papiers filtres contenant du sang ont été prélevés sur une cohorte de 2600 personnes suivie tous les 3 mois jusqu'à 18 mois et analysés par dosage immuno-enzymatique (ELISA). Nos résultats ont montré que les niveaux de réponse IgG à l'antigène gSG6-P1 variaient selon le village, la saison et l'âge et étaient positivement corrélés à l'abondance des espèces anophèles et des vecteurs primaires de paludisme. Une association significative et positive a été observée entre la réponse de l'anticorps au gSG6-P1 et le taux d'inoculation entomologique (EIR), démontrant ainsi que l'hétérogénéité de la transmission du paludisme était directement associée à un comportement de piqûre hétérogène. Des études complémentaires ont montré que le biomarqueur salivaire était pertinent pour détecter des variations micro géographiques dans la transmission à P. falciparum. Cela s’est traduit par des chevauchements significatifs entre les foyers infectieux à P. falciparum et ceux à forts répondeurs en anticorps anti-salive d’anopheles (gSG6-P1). Dans l'ensemble, ces résultats indiquent que le biomarqueur salivaire d'Anopheles est prometteur pour les études épidémiologiques et pourrait guider la mise en œuvre d’interventions de lutte antivectorielle « ciblées » afin d'éliminer les foyers résiduels de paludisme. / Malaria along the Thailand-Myanmar border (TMB) displays geographical heterogeneity and is characterized by high prevalence of submicroscopic carriage and the emergence artemisinin resistance in P. falciparum. Timely identification and elimination of remaining P. falciparum transmission “hotspots” is essential to contain artemisinine resistance. The aim of this study was to address the relevance of using serological biomarkers of human exposure to anopheles bites (gSG6-P1) and Plasmodium antigens to identify remaining sources of transmission and to measure spatial and temporal changes in human vector contact along the TMB. Blood spots were collected in filter papers among a cohort of 2600 people followed every 3 months up to 18 months, and used for analysis by enzyme-linked immunosorbent assay (ELISA). Our findings showed that the levels of IgG responses to gSG6-P1 antigen varied according to village, season, and age and were positively associated with the abundance of total Anopheles species and primary malaria vectors. A significant and positive association was noted between the Antibody response to gSG6-P1 and the entomological inoculation rate (EIR) hence demonstrating that heterogeneity in malaria transmission was directly associated with heterogeneous biting behavior. Further investigations showed that salivary biomarker was relevant to detect small scale variations in P. falciparum malaria. This was supported by scan statistics showing that P. falciparum clusters partially overlap the gSG6-P1 clusters. Altogether, these findings indicates Anopheles salivary biomarker as great potential for epidemiological studies and could be useful to guide the implementation of hotspot–targeted vector control interventions with the aim to achieve malaria elimination.

Die vorliegende Arbeit beschäftigt sich mit der Identifikation neuer immunologischer Indikatoren (Biomarker) für den allgemeinen Gesundheitszustand von Hochleistungsmilchkühen. Diese Biomarker sollen möglichst einfach und schnell mittels eines Stalltests nachweisbar sein, weshalb die gelösten Proteine in der Milch im Fokus standen. Die neuen Biomarker sollten nicht nur Mastitis, sondern vor allem auch Entzündungen außerhalb des Euters anzeigen können. Zu Beginn sollte das Gesamtspektrum an Immunkomponenten erfasst werden, weshalb zunächst auf Proteinexpressionsebene angesetzt wurde. Das schloss die Analyse von vorhandenen Immunzellpopulationen in Blut- und Milchproben ein, um einen Überblick über potentielle Produzenten der immunologischen Indikatoren zu erhalten. Es konnte erstmals Cluster of Differentiation (CD) 25 (alpha-Kette des Interleukin-2-Rezeptors, IL2R) auf bovinen polymorphnukleären, neutrophilen Granulozyten (PMN) aus peripherem Blut nachgewiesen werden. Die Expression (mittlere Fluoreszenzintensität, MFI) von CD25 stieg dabei mit dem Schweregrad der entzündlichen Erkrankung an. Die Ergebnisse konnten auf Transkript- wie auch auf Proteinexpressionsebene bestätigt werden. Gleiche Tendenzen waren auch für Milchzellen erkennbar. In der statistischen Analyse zeigte CD25 auf PMN im peripheren Blut ein hohes Abgrenzungsvermögen für erkrankte Kühe. Die Messung von CD25 auf PMN könnte somit zur Bestimmung des allgemeinen Gesundheitszustandes von Hochleistungsmilchkühen genutzt werden.

BACKGROUND: Distinguishing patients with axial spondyloarthritis (axial SpA) from patients with other causes of chronic back pain remains a challenge. The lack of reliable biomarkers contributes to the diagnostic delay in axial SpA. Recently, macrophage migration inhibitory factor (MIF) has been proposed as a candidate diagnostic and prognostic biomarker. MIF is a proinflammatory cytokine that was shown to be upregulated in several autoimmune diseases, including axial SpA. The putative role of CD8+ T cells in the disease process suggests further that serum markers of cytotoxicity might have value as serological biomarkers in axial SpA, and that subpopulations of cytotoxic lymphocytes might deserve attention as candidate cellular biomarkers. OBJECTIVE: The goal of this study was to compare serum levels of MIF and other candidate serum proteins in patients with axial SpA and controls, and to develop a flow cytometry panel to analyze cytotoxic lymphocyte cell subpopulations in these cohorts, including KIR+CD8+ T cells, Granzyme B+ CD8+ T cells, MAIT cells, and InEx cells. METHODS: Study subjects were recruited from the Brigham and Women’s Hospital Orthopedic and Arthritis Center. Four cohorts were compared: healthy controls (HC), patients with chronic low back pain (cLBP), axial SpA patients not on a biologic (axSpA/-), and axial SpA patients treated with a TNF inhibitor (axSpA/TNFi). Study subjects were matched for age, sex, and race, when possible. Serum was evaluated using the LEGENDplex Human CD8/NK panel (BioLegend) for thirteen markers including IL-17A, IL-6, TNF, granzyme B, and perforin. CRP and MIF were evaluated by DuoSet ELISA (R&D Systems). A high-dimensional flow cytometry panel was designed to evaluate 14 cell populations of interest. RESULTS: The severity of back pain in the cLBP controls and axSpA/- patients was comparable (BASDAI Q2 mean 5.0 +/- 1.9 vs. 5.0 +/- 3.0). axSpA/- patients had higher back pain, BASDAI and ASDAS scores than axSpA/TNFi patients consistent with higher disease activity in the biologic naïve group. Serum CRP values were significantly higher in axSpA/- patients compared with HC, cLBP controls, and axSpA/TNFi patients (P= 0.01, P=0.0029, P=0.004 respectively). Serum MIF levels were not statistically different between all four groups (P= 0.8069). Additionally, there were no statistically significant differences between the groups for any of the markers included in the LEGENDplex Human CD8/NK panel. A 32-color staining panel was developed to evaluate cytotoxic cell populations. CONCLUSION: In contrast to a previous study, we did not find differences in serum MIF levels between axial SpA patients and controls. Of the evaluated serum biomarkers, only CRP values correlated with active axial SpA. We have developed a promising flow cytometry panel that will help analyze subpopulations of cytotoxic cells. This ultimately could shed light on a candidate cellular biomarker. Our results underscore the need for more research into diagnostic biomarkers in axial SpA.

No description available.

An estimated 20.5% of adolescents ages 12 – 19 years were obese (≥95th percentile of BMI-for-age) in 2011 – 2014. Consumption of added sugars (AS) has been linked with adverse effects on weight and cardiovascular disease risk factors. Approximately 16% of adolescents’ calories come from AS, of which sugar-sweetened beverages (SSB) are a major contributor. However, the relationship between AS/SSB intake and obesity is controversial, partly due to limitations in self-reported dietary data. Objective dietary intake biomarkers may circumvent this problem. The δ13C biomarker for AS intake is based upon the fact that C4 plants– major source for sugar production in the United States – have elevated δ¹³C values compared to C3 plants, which includes most fruits and vegetables. The δ¹³C value of blood, which is influenced by diet, has been established as a valid, reliable, and sensitive biomarker, but when compared to selfreported AS intake. This investigation evaluated the sensitivity and reliability of the δ13C biomarker, assessed with fingerstick blood samples, in adolescents using a controlled feeding, crossover design. Fingerstick δ¹³C values significantly changed by -0.05‰ and +0.03‰ after subjects completed the 5% and 25% AS diets, respectively (F(1, 30) = 18.828, p < 0.001). High reliability was found between two consecutive fingerstick δ¹³C values on the low (ICC = 0.996) and high (ICC = 0.997) AS diets. Thus, fingerstick δ¹³C may be a sensitive and reliable indicator of AS intake in adolescents. Future investigations should develop an equation to estimate AS intake based on fingerstick δ¹³C / Master of Science / Approximately one-fifth of adolescents 12 – 19 years old were obese in 2011 – 2014. A diet high in added sugars (AS), which are sugars that do not naturally occur in a food product, is associated with increased weight and higher risk for cardiovascular, or heart, disease. About 16% of adolescents’ calories come from AS, and a major source of AS intake is sugary beverages. Because people tend to inaccurately report what they eat and drink, researchers are interested in biomarkers to objectively estimate dietary intake. The δ¹³C biomarker measures carbon isotope ratios. Since most of the sugar produced in the United States comes from corn, sugarcane, and sorghum – which have a higher δ ¹³C content compared to most fruits and vegetables – δ ¹³C could indicate AS intake. Studies have reported that the δ ¹³C value of whole blood, which is influenced by diet, is valid, reliable, and sensitive, but when compared to self-reported AS intake. This investigation evaluated the sensitivity and reliability of whole blood δ ¹³C, sample using fingersticks, in adolescents consuming controlled diets so that AS intake was known. Fingerstick δ ¹³C values significantly changed after subjects completed the low and high AS diets (<i>F</i>(1, 30) = 18.828, <i>p</i> < 0.001). High reliability was found between two consecutive fingerstick δ ¹³C values on the low (ICC = 0.996) and high (ICC = 0.997) AS diets. Thus, fingerstick δ ¹³C may be a sensitive, reliable indicator of AS intake in adolescents. Future studies should develop an equation to estimate AS intake based on fingerstick δ ¹³C.

abstract: Urinary sucrose and fructose has been suggested as a predictive biomarker of total sugars intake based on research involving UK adults. The purpose of this study was to determine the association between total sugars consumption and 24-hour urinary sucrose and fructose (24uSF) in US adult population and to investigate the effect of physical activity on this association. Fifty seven free-living healthy subjects 20 to 68 years old, participated in a 15-day highly controlled feeding study, consuming their habitual diet, provided by the research metabolic kitchen. Dietary sugars were estimated using Nutrition Data System for Research (NDSR). Subjects collected eight 24-hour urine samples measured for urinary sucrose and fructose. Physical activity was assessed daily using a validated 15-day log that inquired about 38 physical activities across six domains; home activities, transportation, occupation, conditioning, sports and leisure. The mean total sugars intake and added sugars intake of the sample was 112.2 (33.1) g/day and 65.8 (29.0) g/day (9.7%EI), respectively. Significant moderate positive correlation was found between 15-d mean total sugars intake and 8-day mean 24uSF (r = 0.56, p < 0.001). Similarly, added sugars were moderately correlated with 24uSF (r = 0.56, p < 0.001), while no correlation was found between naturally-occurring sugars and 24uSF (r = 0.070, p < 0.001). In a linear multiple regression, total and added sugars each explained 30% of variability in 24uSF (Adjusted R2, p value; total sugars: 0.297, 0.001; added sugars: 0.301, p < 0.001). Physical activity had no effect on the association between dietary and urinary sugars in neither the correlation nor the linear regression analysis. 24uSF can be used as a biomarker for total and added sugars consumption in US adults, although its predictability was weaker compared to findings involving UK adults. No evidence was found showing that physical activity levels affect the association between 24uSF and total sugars intake in US adults. More detailed investigation through future feeding studies including subjects with wide range of sugars intake and of different ethnic/racial backgrounds are needed to better understand the characteristics of the biomarker and its uses. / Dissertation/Thesis / Masters Thesis Nutrition 2019

Mitochondria are the main site of cellular adenosine triphosphate (ATP) generation which is achieved by a series of multi-subunit complexes and electron carriers which together create the oxidative phosphorylation system (OXPHOS). Whenever a defect in any of the numerous mitochondrial pathways occurs it is commonly referred to as a mitochondrial disorder. Mitochondrial disorders are a heterogeneous group of disorders characterised by impaired energy production and include a wide range of defects of either mitochondrial DNA (mtDNA) or nuclear DNA (nDNA) encoded proteins. In cases of dysfunction in the respiratory chain (complex I to IV) it is known to be a respiratory chain deficiency (RCD) which presents a huge challenge for routine diagnosis largely due to the lack of a specific and sensitive biomarker(s). One sure way of confirming the suspicion of a RCD is by performing enzyme analysis on a muscle sample obtained through a biopsy. However, due to the lack of theatre time available to clinicians and the relative large number of false positive patients that are being selected for biopsies, it was decided to develop a biosignature to limit the number of false positive patients from the diagnostic workflow. An untargeted liquid chromatography mass spectrometry (LC-MS) metabolomics approach was used to investigate RCDs in children from South Africa. Sample preparation, a liquid chromatography time-of-flight mass spectrometry method and data processing methods were standardised. Furthermore the developed methodology made use of reverse phase chromatography in conjunction with positive electrospray ionisation (ESI) and a hydrophilic interaction chromatography (HILIC) in negative electrospray ionisation. Urine samples of 61 patients representing three different experimental groups were analysed. The three experimental groups comprised of patients with respiratory chain deficiencies, clinical referred controls (CRC) and patients suffering from various neuromuscular disorders (NMD). After a variety of data mining steps and statistical analysis a list of 12 features were compiled with the ability to distinguish between patients with RCDs and CRCs. The proposed signature was also tested on the neuromuscular disorder group, but this result indicated that the biosignature performed better when used to differentiate between patients with RCDs and CRCs, since the model was designed with this purpose. An alternative validation study is required to identify the features found with this proposed biosignature, to ensure that this biosignature can be practically implemented as a non-invasive screening method. / MSc (Chemistry), North-West University, Potchefstroom Campus, 2014

Mitochondria are the main site of cellular adenosine triphosphate (ATP) generation which is achieved by a series of multi-subunit complexes and electron carriers which together create the oxidative phosphorylation system (OXPHOS). Whenever a defect in any of the numerous mitochondrial pathways occurs it is commonly referred to as a mitochondrial disorder. Mitochondrial disorders are a heterogeneous group of disorders characterised by impaired energy production and include a wide range of defects of either mitochondrial DNA (mtDNA) or nuclear DNA (nDNA) encoded proteins. In cases of dysfunction in the respiratory chain (complex I to IV) it is known to be a respiratory chain deficiency (RCD) which presents a huge challenge for routine diagnosis largely due to the lack of a specific and sensitive biomarker(s). One sure way of confirming the suspicion of a RCD is by performing enzyme analysis on a muscle sample obtained through a biopsy. However, due to the lack of theatre time available to clinicians and the relative large number of false positive patients that are being selected for biopsies, it was decided to develop a biosignature to limit the number of false positive patients from the diagnostic workflow. An untargeted liquid chromatography mass spectrometry (LC-MS) metabolomics approach was used to investigate RCDs in children from South Africa. Sample preparation, a liquid chromatography time-of-flight mass spectrometry method and data processing methods were standardised. Furthermore the developed methodology made use of reverse phase chromatography in conjunction with positive electrospray ionisation (ESI) and a hydrophilic interaction chromatography (HILIC) in negative electrospray ionisation. Urine samples of 61 patients representing three different experimental groups were analysed. The three experimental groups comprised of patients with respiratory chain deficiencies, clinical referred controls (CRC) and patients suffering from various neuromuscular disorders (NMD). After a variety of data mining steps and statistical analysis a list of 12 features were compiled with the ability to distinguish between patients with RCDs and CRCs. The proposed signature was also tested on the neuromuscular disorder group, but this result indicated that the biosignature performed better when used to differentiate between patients with RCDs and CRCs, since the model was designed with this purpose. An alternative validation study is required to identify the features found with this proposed biosignature, to ensure that this biosignature can be practically implemented as a non-invasive screening method. / MSc (Chemistry), North-West University, Potchefstroom Campus, 2014