Learning Methods and Algorithms for Semantic Text Classification across Multiple Domains

Information is nowadays a key resource: machine learning and data mining techniques have been developed to extract high-level information from great amounts of data. As most data comes in form of unstructured text in natural languages, research on text mining is currently very active and dealing with practical problems. Among these, text categorization deals with the automatic organization of large quantities of documents in priorly defined taxonomies of topic categories, possibly arranged in large hierarchies. In commonly proposed machine learning approaches, classifiers are automatically trained from pre-labeled documents: they can perform very accurate classification, but often require a consistent training set and notable computational effort. Methods for cross-domain text categorization have been proposed, allowing to leverage a set of labeled documents of one domain to classify those of another one. Most methods use advanced statistical techniques, usually involving tuning of parameters. A first contribution presented here is a method based on nearest centroid classification, where profiles of categories are generated from the known domain and then iteratively adapted to the unknown one. Despite being conceptually simple and having easily tuned parameters, this method achieves state-of-the-art accuracy in most benchmark datasets with fast running times. A second, deeper contribution involves the design of a domain-independent model to distinguish the degree and type of relatedness between arbitrary documents and topics, inferred from the different types of semantic relationships between respective representative words, identified by specific search algorithms. The application of this model is tested on both flat and hierarchical text categorization, where it potentially allows the efficient addition of new categories during classification. Results show that classification accuracy still requires improvements, but models generated from one domain are shown to be effectively able to be reused in a different one.

Spectral theory of differential operators on finite metric graphs and on bounded domains

Die vorliegende Arbeit widmet sich der Spektraltheorie von Differentialoperatoren auf metrischen Graphen und von indefiniten Differentialoperatoren auf beschränkten Gebieten. Sie besteht aus zwei Teilen. Im Ersten werden endliche, nicht notwendigerweise kompakte, metrische Graphen und die Hilberträume von quadratintegrierbaren Funktionen auf diesen betrachtet. Alle quasi-m-akkretiven Laplaceoperatoren auf solchen Graphen werden charakterisiert, und Abschätzungen an die negativen Eigenwerte selbstadjungierter Laplaceoperatoren werden hergeleitet. Weiterhin wird die Wohlgestelltheit eines gemischten Diffusions- und Transportproblems auf kompakten Graphen durch die Anwendung von Halbgruppenmethoden untersucht. Eine Verallgemeinerung des indefiniten Operators $-tfrac{d}{dx}sgn(x)tfrac{d}{dx}$ von Intervallen auf metrische Graphen wird eingeführt. Die Spektral- und Streutheorie der selbstadjungierten Realisierungen wird detailliert besprochen. Im zweiten Teil der Arbeit werden Operatoren untersucht, die mit indefiniten Formen der Art $langlegrad v, A(cdot)grad urangle$ mit $u,vin H_0^1(Omega)subset L^2(Omega)$ und $OmegasubsetR^d$ beschränkt, assoziiert sind. Das Eigenwertverhalten entspricht in Dimension $d=1$ einer verallgemeinerten Weylschen Asymptotik und für $dgeq 2$ werden Abschätzungen an die Eigenwerte bewiesen. Die Frage, wann indefinite Formmethoden für Dimensionen $dgeq 2$ anwendbar sind, bleibt offen und wird diskutiert.

Das GAF-Domänen-Protein NreA : ein neuartiger Nitratrezeptor aus Staphylococcus carnosus

Staphylococcus carnosus ist ein fakultativ anaerobes Bakterium, das aerobe Atmung, anaerobe Nitratatmung und Gärungsstoffwechsel betreiben kann. Die Expression des Nitratstoffwechsels wird durch das Dreikomponentensystem NreABC reguliert.rnUnter anaeroben Bedingungen besitzt die Sensorhistidinkinase NreB in ihrer PAS-Domäne ein [Fe4S4]2+-Cluster. Das aktive (anaerobe) [Fe4S4]2+-NreB überträgt nach Autophosphorylierung die Phosphorylgruppe auf den Antwortregulator NreC, welcher dann die Expression der Gene der Nitratatmung aktiviert. Nitrat wirkt mit Hilfe des NreA-Proteins auf diese Gene induzierend. Im Rahmen der vorliegenden Arbeit wurde gezeigt, dass NreA ein GAF-Domänen-Protein und ein neuartiger Nitratrezeptor ist.rnDie Natur von NreA als GAF-Domänen-Protein bestätigte sich beim Vergleich der Kristallstruktur mit denen anderer GAF-Domänen. GAF-Domänen sind weit verbreitet und binden typischer Weise kleine Moleküle. Als physiologischer Ligand von NreA zeigte sich Nitrat, das innerhalb einer definierten Bindetasche gebunden wird. NreA bindet vermutlich in dimerer Form an dimeres NreB und inhibiert dadurch die Phosphorylierung der Sensorhistidinkinase NreB. Die Interaktion von NreA mit NreB wurde in vivo durch BACTH-Messungen und sowohl in vivo als auch in vitro durch Cross-Linking Experimente gezeigt. Nitrat reduziert den Ergebnissen nach die Interaktion von NreA mit NreB.rnDurch Sequenzvergleiche von NreA mit Homologen wurden konservierte Aminosäuren identifiziert. Über gerichtete Mutagenese wurden 25 NreA-Varianten hergestellt und bezüglich ihres Verhaltens in Abhängigkeit von Nitrat in narG-lip-Reportergenstudien getestet. Anhand ihres Phänotyps wurden sie als Wildtyp, NreA- und NreABC-Mutanten klassifiziert. Die Nitratbindetasche war in sechs Fällen betroffen. Die Phänotypen der Mutationen in der Peripherie lassen sich mit Auswirkungen auf die vermutete Konformationsänderung oder auf die Interaktion mit NreB erklären. Mutationen von konservierten, oberflächenexponierten Resten führten vermehrt zu NreA/ON-Varianten. Es ließen sich Bereiche auf der Proteinoberfläche identifizieren, die für NreA/NreA- oder NreA/NreB-Interaktionen wichtig sein könnten.rnDie Untersuchungen zeigten, dass NreA mit NreB interagiert und dass dadurch ein NreA/NreB-Sensorkomplex für die gemeinsame Erkennung von Nitrat und Sauerstoff gebildet wird.

Welche Typen von Autokratien sind leistungsfähiger und warum? : eine empirische Analyse auf Grundlage der Selektoratstheorie = Which type of autocracy performs better, and why?

Performanzunterschiede zwischen verschiedenen Autokratietypen wie Monarchien, Militär-, Ein-Parteien- und begrenzten Mehr-Parteien-Regimen sind bis auf wenige Ausnahmen bisher relativ unerforscht. Deshalb widmet sich diese Arbeit folgenden Forschungsfragen: Gibt es Performanzunterschiede zwischen verschiedenen Autokratietypen? Wenn ja, wie gestalten sich diese und wie lassen sie sich erklären? Auf Grundlage der Selektoratstheorie von Bueno de Mesquite et al. wird die Hypothese aufgestellt, dass die Performanz politischer Regime mit der Größe der Winning Coalition steigt. Da verschiedene Autokratietypen unterschiedlich große Winning Coalitions haben, wird angenommen, dass es deutliche Performanzunterschiede zwischen diesen Typen gibt. Als Performanzkriterien dienen in dieser Arbeit wirtschaftlicher Wohlstand, soziale Sicherheit und ökologische Nachhaltigkeit. Aus diesen drei Kriterien wird zusätzlich ein Indikator allgemeiner Performanz gebildet. Die empirische Untersuchung erfolgt mit den Daten des Quality-of-Governement-Datensatzes und erstreckt sich über 140 autokratische Länder im Zeitraum von 1972 bis 2010. Die Daten werden mittels Time-Series-Cross-Section-Regressionen analysiert. Die Ergebnisse der Analysen entsprechen nur teilweise den Erwartungen. Die Größe der Winning Coalition wirkt positiv auf die wirtschaftliche Wohlfahrt und die soziale Sicherheit und damit auch auf die allgemeine Performanz aus. Entgegen den Erwartungen sinkt die ökologische Performanz jedoch mit steigender Größe der Winning Coalition. Auch die Befunde bezüglich der Performanz verschiedener Autokratietypen entsprechen nicht den Erwartungen. So sind Ein-Parteien-Regime insgesamt leistungsfähiger als Mehr-Parteien-Regime, Militär-Regime und Monarchien. Militär-Regime sind leistungsfähiger als Monarchien und tendenziell auch als Mehr-Parteien-Regime.

TBX3-Mutationsanalysen und Konstruktion einer Zelllinie für TBX2-Inhibitorscreenings

Die nahe verwandten T-box Transkriptionsfaktoren TBX2 und TBX3 werden in zahlreichen humanen Krebsarten überexprimiert, insbesondere in Brustkrebs und Melanomen. Die Überexpression von TBX2 und TBX3 hat verschiedene zelluläre Effekte, darunter die Unterdrückung der Seneszenz, die Förderung der Epithelialen-Mesenchymalen Transition sowie invasive Zellmotilität. Im Gegensatz dazu führt ein Funktionsverlust von TBX3 und der meisten anderen humanen T-box-Gene zu haploinsuffizienten Entwicklungsdefekten. Durch Sequenzierung des Exoms von Brustkrebsproben identifizierten Stephens et al. fünf verschiedene Mutationen in TBX3, welche allesamt die DNA-bindende T-box-Domäne betrafen. Die In-Frame-Deletion N212delN wurde zweimal gefunden. Aus der Anhäufung der Mutationen innerhalb der T-box-Domäne wurde geschlossen, dass TBX3 bei Brustkrebs ein Treibergen ist. Da Mutationen innerhalb der T-box-Domäne im Allgemeinen zu einem Funktionsverlust führen, aber die onkogene Aktivität von TBX3 meist auf eine Überexpression zurückzuführen ist, wurden die potentiellen Treibermutationen hinsichtlich einer verminderten oder gesteigerten TBX3-Funktion geprüft. Getestet wurden zwei In-Frame Deletionen, eine Missense- sowie eine Frameshift-Mutante bezüglich der DNA-Bindung in vitro und der Zielgen-Repression in Zellkultur. Zusätzlich wurde eine in silico Analyse der im The Cancer Genome Atlas (TCGA) gelisteten somatischen TBX-Brustkrebsmutationen durchgeführt. Sowohl die experimentelle als auch die in silico Analyse zeigten, dass die untersuchten Mutationen vorwiegend zum Verlust der TBX3-Funktion führen. Um den Mechanismus der Genrepression durch TBX3 besser zu verstehen, wurden weitere TBX3-Mutanten bezüglich ihrer Wirkung auf die p21-Promotoraktivität (p21-Luc-Reporter und endogene p21-Expression) analysiert. Wildtypische p21-Luc-Repression zeigten die zwei Mutationen S674A (Phosphorylierung) und D275K (SUMOylierung), welche posttranslationale Modifikationen verhindern, sowie die Interaktion mit dem Tumorsuppressor Rb1 unterbindende M302A/V304A-Mutation. Erstaunlicherweise war die endogene p21-Repression dieser Mutanten stärker als die des wildtypischen TBX3-Proteins. Alle drei Mutationen führten zu einer Stabilisierung des TBX3-Proteins. Die ursprünglich in Patienten mit Ulna-Mamma Syndrom identifizierte, DNA-bindungsdefekte Y149S-Mutante konnte weder p21-Luc noch endogenes p21 reprimieren. Mutationen in potentiellen Interaktionsdomänen für die Bindung der Co-Repressoren Groucho und C-terminalem Bindeprotein zeigten sowohl auf p21-Luc als auch auf endogenes p21-Gen wildtypische Repressoraktivität, so dass diese Co-Repressoren in COS-7-Zellen wahrscheinlich nicht an der Repression dieses Gens beteiligt sind. Da TBX2 und TBX3 interessante Ziele zur direkten Krebsbekämpfung darstellen, sollte ein zelluläres Reportersystem zur Identifikation TBX2-inhibierender, pharmakologisch aktiver Substanzen etabliert werden. Dazu sollte eine stabile Zelllinie mit vom p21-Promotor reguliertem d2EGFP-Reporter und Doxyzyklin-induzierbarem TBX2-Protein erzeugt werden, da ektopische Expression von TBX2 genetische Instabilität und Toxizität induzieren kann. In dieser Zelllinie sollte die TBX2-Expression zur Reduktion der d2EGFP-Fluoreszenz führen. Zur Erzeugung der Zelllinie wurden die folgenden drei Konstrukte Schritt-für-Schritt stabil in das Genom der Zielzelllinie COS-7 integriert: pEF1alpha-Tet3G, pTRE3G-TBX2 und p21-d2EGFP. Während die Herstellung der doppelt stabilen COS-7-Zelllinie gelang, scheiterte die Herstellung der dreifach stabilen Zelllinie.

Older Breast Cancer Survivors: Geriatric Assessment Domains Are Associated With Poor Tolerance of Treatment Adverse Effects and Predict Mortality Over 7 Years of Follow-Up

Purpose To evaluate geriatric assessment (GA) domains in relation to clinically important outcomes in older breast cancer survivors. Methods Six hundred sixty women diagnosed with primary breast cancer in four US geographic regions (Los Angeles, CA; Minnesota; North Carolina; and Rhode Island) were selected with disease stage I to IIIA, age ≥ 65 years at date of diagnosis, and permission from attending physician to contact. Data were collected over 7 years of follow-up from consenting patients' medical records, telephone interviews, physician questionnaires, and the National Death Index. Outcomes included self-reported treatment tolerance and all-cause mortality. Four GA domains were described by six individual measures, as follows: sociodemographic by adequate finances; clinical by Charlson comorbidity index (CCI) and body mass index; function by number of physical function limitations; and psychosocial by the five-item Mental Health Index (MHI5) and Medical Outcomes Study Social Support Survey (MOS-SSS). Associations were evaluated using t tests, χ2 tests, and regression analyses. Results In multivariable regression including age and stage, three measures from two domains (clinical and psychosocial) were associated with poor treatment tolerance; these were CCI ≥ 1 (odds ratio [OR] = 2.49; 95% CI, 1.18 to 5.25), MHI5 score less than 80 (OR = 2.36; 95% CI, 1.15 to 4.86), and MOS-SSS score less than 80 (OR = 3.32; 95% CI, 1.44 to 7.66). Four measures representing all four GA domains predicted mortality; these were inadequate finances (hazard ratio [HR] = 1.89; 95% CI, 1.24 to 2.88; CCI ≥ 1 (HR = 1.38; 95% CI, 1.01 to 1.88), functional limitation (HR = 1.40; 95% CI, 1.01 to 1.93), and MHI5 score less than 80 (HR = 1.34; 95% CI, 1.01 to 1.85). In addition, the proportion of women with these outcomes incrementally increased as the number of GA deficits increased. Conclusion This study provides longitudinal evidence that GA domains are associated with poor treatment tolerance and predict mortality at 7 years of follow-up, independent of age and stage of disease.

Oxidative stress and inflammation response after nanoparticle exposure: differences between human lung cell monocultures and an advanced three-dimensional model of the human epithelial airways

Combustion-derived and manufactured nanoparticles (NPs) are known to provoke oxidative stress and inflammatory responses in human lung cells; therefore, they play an important role during the development of adverse health effects. As the lungs are composed of more than 40 different cell types, it is of particular interest to perform toxicological studies with co-cultures systems, rather than with monocultures of only one cell type, to gain a better understanding of complex cellular reactions upon exposure to toxic substances. Monocultures of A549 human epithelial lung cells, human monocyte-derived macrophages and monocyte-derived dendritic cells (MDDCs) as well as triple cell co-cultures consisting of all three cell types were exposed to combustion-derived NPs (diesel exhaust particles) and to manufactured NPs (titanium dioxide and single-walled carbon nanotubes). The penetration of particles into cells was analysed by transmission electron microscopy. The amount of intracellular reactive oxygen species (ROS), the total antioxidant capacity (TAC) and the production of tumour necrosis factor (TNF)-alpha and interleukin (IL)-8 were quantified. The results of the monocultures were summed with an adjustment for the number of each single cell type in the triple cell co-culture. All three particle types were found in all cell and culture types. The production of ROS was induced by all particle types in all cell cultures except in monocultures of MDDCs. The TAC and the (pro-)inflammatory reactions were not statistically significantly increased by particle exposure in any of the cell cultures. Interestingly, in the triple cell co-cultures, the TAC and IL-8 concentrations were lower and the TNF-alpha concentrations were higher than the expected values calculated from the monocultures. The interplay of different lung cell types seems to substantially modulate the oxidative stress and the inflammatory responses after NP exposure.

Long-term follow-up of patients with Bartter syndrome type I and II

BACKGROUND: Little information is available on a long-term follow-up in Bartter syndrome type I and II. METHODS: Clinical presentation, treatment and long-term follow-up (5.0-21, median 11 years) were evaluated in 15 Italian patients with homozygous (n = 7) or compound heterozygous (n = 8) mutations in the SLC12A1 (n = 10) or KCNJ1 (n = 5) genes. RESULTS: Thirteen new mutations were identified. The 15 children were born pre-term with a normal for gestational age body weight. Medical treatment at the last follow-up control included supplementation with potassium in 13, non-steroidal anti-inflammatory agents in 12 and gastroprotective drugs in five patients. At last follow-up, body weight and height were within normal ranges in the patients. Glomerular filtration rate was <90 mL/min/1.73 m(2) in four patients (one of them with a pathologically increased urinary protein excretion). In three patients, abdominal ultrasound detected gallstones. The group of patients with antenatal Bartter syndrome had a lower renin ratio (P < 0.05) and a higher standard deviation score (SDS) for height (P < 0.05) than a previously studied group of patients with classical Bartter syndrome. CONCLUSIONS: Patients with Bartter syndrome type I and II tend to present a satisfactory prognosis after a median follow-up of more than 10 years. Gallstones might represent a new complication of antenatal Bartter syndrome.

Computed tomographic coronary angiography in patients with surgically treated type A aortic dissection: preliminary results

Acute type A aortic dissection is a serious emergency with a mortality rate of up to 40% within the first 24 h when left untreated. Surgical therapy needs to be initiated promptly. Due to this urgent situation, preoperative evaluation of the coronary arteries is not routinely performed in these patients. The aim of this study was to evaluate the accuracy of 64-slice computed tomography angiography (CTA) for postoperative coronary artery assessment in these patients. Ten consecutive patients with two or more cardiovascular risk factors were prospectively enrolled. Patients had type A aortic dissection treated surgically with a supracoronary graft of the ascending aorta. Performance of CTA to exclude significant stenosis (>50% lumen narrowing) and/or coronary artery dissection was compared with quantitative coronary angiography. A total of 147 segments were evaluated. Three segments (2%) were excluded from analysis. CTA correctly assessed one of three significant stenoses in three patients and correctly excluded coronary artery disease (CAD) in six of ten patients. One patient was rated false positive. Overall accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of CT for identifying coronary artery disease by segment was 98%, 33%, 99%, 50%, and 99%, respectively (P<0.05). By patient, it was 70%, 33%, 86%, 50%, and 75%, respectively. No coronary artery dissection was found. Noninvasive CTA may be a viable alternative to conventional angiography for postoperative coronary artery evaluation in patients with surgically treated type A aortic dissection and cardiovascular risk factors. An NPV of 99% should allow for reliable exclusion of CAD. Further studies with higher patient numbers are warranted.

Ramachandran-type Plots for Glycosidic Linkages: Examples from Molecular Dynamics Simulations using the Glycam06 Force Field

The goals of this article are to (1) provide further validation of the Glycam06 force field, specifically for its use in implicit solvent molecular dynamic (MD) simulations, and (2) to present the extension of G.N. Ramachandran's idea of plotting amino acid phi and psi angles to the glycosidic phi, psi, and omega angles formed between carbohydrates. As in traditional Ramachandran plots, these carbohydrate Ramachandran-type (carb-Rama) plots reveal the coupling between the glycosidic angles by displaying the allowed and disallowed conformational space. Considering two-bond glycosidic linkages, there are 18 possible conformational regions that can be defined by (α, ϕ, ψ) and (β, ϕ, ψ), whereas for three-bond linkages, there are 54 possible regions that can be defined by (α, ϕ, ψ, ω) and (β, ϕ, ψ, ω). Illustrating these ideas are molecular dynamic simulations on an implicitly hydrated oligosaccharide (700 ns) and its eight constituent disaccharides (50 ns/disaccharide). For each linkage, we compare and contrast the oligosaccharide and respective disaccharide carb-Rama plots, validate the simulations and the Glycam06 force field through comparison to experimental data, and discuss the general trends observed in the plots.

Chapter VI: Follow-up after revascularisation

Structured follow-up after revascularisation for chronic critical limb ischaemia (CLI) aims at sustained treatment success and continued best patient care. Thereby, efforts need to address three fundamental domains: (A) best medical therapy, both to protect the arterial reconstruction locally and to reduce atherosclerotic burden systemically; (B) surveillance of the arterial reconstruction; and (C) timely initiation of repeat interventions. As most CLI patients are elderly and frail, sustained resolution of CLI and preserved ambulatory capacity may decide over independent living and overall prognosis. Despite this importance, previous guidelines have largely ignored follow-up after CLI; arguably because of a striking lack of evidence and because of a widespread assumption that, in the context of CLI, efficacy of initial revascularisation will determine prognosis during the short remaining life expectancy. This chapter of the current CLI guidelines aims to challenge this disposition and to recommend evidentially best clinical practice by critically appraising available evidence in all of the above domains, including antiplatelet and antithrombotic therapy, clinical surveillance, use of duplex ultrasound, and indications for and preferred type of repeat interventions for failing and failed reconstructions. However, as corresponding studies are rarely performed among CLI patients specifically, evidence has to be consulted that derives from expanded patient populations. Therefore, most recommendations are based on extrapolations or subgroup analyses, which leads to an almost systematic degradation of their strength. Endovascular reconstruction and surgical bypass are considered separately, as are specific contexts such as diabetes or renal failure; and critical issues are highlighted throughout to inform future studies.

Influence of V5/6-His Tag on the Properties of Gap Junction Channels Composed of Connexin43, Connexin40 or Connexin45

HeLa cells expressing wild-type connexin43, connexin40 or connexin45 and connexins fused with a V5/6-His tag to the carboxyl terminus (CT) domain (Cx43-tag, Cx40-tag, Cx45-tag) were used to study connexin expression and the electrical properties of gap junction channels. Immunoblots and immunolabeling indicated that tagged connexins are synthesized and targeted to gap junctions in a similar manner to their wild-type counterparts. Voltage-clamp experiments on cell pairs revealed that tagged connexins form functional channels. Comparison of multichannel and single-channel conductances indicates that tagging reduces the number of operational channels, implying interference with hemichannel trafficking, docking and/or channel opening. Tagging provoked connexin-specific effects on multichannel and single-channel properties. The Cx43-tag was most affected and the Cx45-tag, least. The modifications included (1) V j-sensitive gating of I j (V j, gap junction voltage; I j, gap junction current), (2) contribution and (3) kinetics of I j deactivation and (4) single-channel conductance. The first three reflect alterations of fast V j gating. Hence, they may be caused by structural and/or electrical changes on the CT that interact with domains of the amino terminus and cytoplasmic loop. The fourth reflects alterations of the ion-conducting pathway. Conceivably, mutations at sites remote from the channel pore, e.g., 6-His-tagged CT, affect protein conformation and thus modify channel properties indirectly. Hence, V5/6-His tagging of connexins is a useful tool for expression studies in vivo. However, it should not be ignored that it introduces connexin-dependent changes in both expression level and electrophysiological properties.

Redox-Switching in a Viologen-type Adlayer: An Electrochemical Shell-Isolated Nanoparticle Enhanced Raman Spectroscopy Study on Au(111)-(1 x 1) Single Crystal Electrodes

We reported the first application of in situ shell-isolated nanoparticle enhanced Raman spectroscopy (SHINERS) to an interfacial redox reaction under electrochemical conditions. We construct gap-mode sandwich structures composed of a thiol-terminated HS-6V6H viologen adlayer immobilized on a single crystal Au(111)-(1x1) electrode and covered by Au(60 nm)@SlO(2) core shell nanoparticles acting as plasmonic antennas. We observed high-quality, potential-dependent Raman spectra of the three viologen species V(2+),V(+center dot) and V(0) on a well-defined Au(111) substrate surface and could map their potential-dependent evolution. Comparison with experiments on powder samples revealed an enhancement factor of the nonresonant Raman modes of similar to 3 x 10(5), and up to 9 x 10(7) for the resonance modes. The study illustrates the unique capability of SHINERS and its potential in the entire field of electrochemical surface science to explore structures and reaction pathways on well-defined substrate surfaces, such as single crystals, for molecular, (electro-)- catalytic, bioelectrochemical systems up to fundamental double layer studies at electrified solid/liquid interfaces.

Influence of time point of calibration on accuracy of continuous glucose monitoring in individuals with type 1 diabetes

Abstract Background and Aims: Data on the influence of calibration on accuracy of continuous glucose monitoring (CGM) are scarce. The aim of the present study was to investigate whether the time point of calibration has an influence on sensor accuracy and whether this effect differs according to glycemic level. Subjects and Methods: Two CGM sensors were inserted simultaneously in the abdomen on either side of 20 individuals with type 1 diabetes. One sensor was calibrated predominantly using preprandial glucose (calibration(PRE)). The other sensor was calibrated predominantly using postprandial glucose (calibration(POST)). At minimum three additional glucose values per day were obtained for analysis of accuracy. Sensor readings were divided into four categories according to the glycemic range of the reference values (low, ≤4 mmol/L; euglycemic, 4.1-7 mmol/L; hyperglycemic I, 7.1-14 mmol/L; and hyperglycemic II, >14 mmol/L). Results: The overall mean±SEM absolute relative difference (MARD) between capillary reference values and sensor readings was 18.3±0.8% for calibration(PRE) and 21.9±1.2% for calibration(POST) (P<0.001). MARD according to glycemic range was 47.4±6.5% (low), 17.4±1.3% (euglycemic), 15.0±0.8% (hyperglycemic I), and 17.7±1.9% (hyperglycemic II) for calibration(PRE) and 67.5±9.5% (low), 24.2±1.8% (euglycemic), 15.5±0.9% (hyperglycemic I), and 15.3±1.9% (hyperglycemic II) for calibration(POST). In the low and euglycemic ranges MARD was significantly lower in calibration(PRE) compared with calibration(POST) (P=0.007 and P<0.001, respectively). Conclusions: Sensor calibration predominantly based on preprandial glucose resulted in a significantly higher overall sensor accuracy compared with a predominantly postprandial calibration. The difference was most pronounced in the hypo- and euglycemic reference range, whereas both calibration patterns were comparable in the hyperglycemic range.

An Actor-Critic based controller for glucose regulation in type 1 diabetes

A novel adaptive approach for glucose control in individuals with type 1 diabetes under sensor-augmented pump therapy is proposed. The controller, is based on Actor-Critic (AC) learning and is inspired by the principles of reinforcement learning and optimal control theory. The main characteristics of the proposed controller are (i) simultaneous adjustment of both the insulin basal rate and the bolus dose, (ii) initialization based on clinical procedures, and (iii) real-time personalization. The effectiveness of the proposed algorithm in terms of glycemic control has been investigated in silico in adults, adolescents and children under open-loop and closed-loop approaches, using announced meals with uncertainties in the order of ±25% in the estimation of carbohydrates. The results show that glucose regulation is efficient in all three groups of patients, even with uncertainties in the level of carbohydrates in the meal. The percentages in the A+B zones of the Control Variability Grid Analysis (CVGA) were 100% for adults, and 93% for both adolescents and children. The AC based controller seems to be a promising approach for the automatic adjustment of insulin infusion in order to improve glycemic control. After optimization of the algorithm, the controller will be tested in a clinical trial.