Search

Biochemical systems have been influenced, altered, and engineered to produce a myriad of complex behaviours that have far reaching consequences in the realization of computational models and in applications such as pharmaceuticals and cell networks. However, unlike producing desired behaviours in digital systems, the mechanics behind biochemical systems are complex and poorly understood. In particular, gaps in our understanding of the behaviour of cell mechanics or synthetic molecular systems have hindered the production of biochemical circuits that model or enact certain behaviours. This thesis addresses these issues by accepting the inherent uncertainty and unknowns within biochemical systems, and exploring ways to produce desired behaviours despite these difficulties. This is done through the use of deterministic and stochastic interpretations of Chemical Reaction Networks (CRNs) to describe these biochemical systems. The first part of this thesis considers parameter synthesis of CRNs. A novel sketching language for CRNs that allows for both discrete and continuous parameter declarations is proposed. Often it is not only the correctness of synthesized programs that is important, but also their optimality with respect to a given cost function. Based on a cost function given by the structure of a CRN, there is an attempt to reduce the cost in order to produce a CRN optimal for the given cost function. Synthesis of complex CRN behaviour without structural constraints is seen as intractable, and therefore Syntax-Guided Synthesis (SyGuS) is employed to constrain the search space and allow us to synthesize behaviour of non-linear Ordinary Differential Equations (ODEs). Algorithms and case studies aimed at finding an optimal CRN are provided. Satisfiability Modulo Theory (SMT-ODE) solvers are employed to solve parametric ODEs constructed from a combination of the Linear Noise Approximation dynamics and sketching language choice variables. In our tool, named CRNSketch, a generalization of the CRN synthesis problem to include non mass-action kinetics as well as arbitrary continuous functions as inputs to a CRN, is successfully demonstrated. Biologically motivated case studies are provided highlighting the need for the tool, as well as our parameter and structural synthesis methods in general. An evaluation is provided on the limitations and tractability of our parameter synthesis approach. The second part of this thesis addresses the problem of control of CRNs, because of the inherent unknowns in biochemical systems. Using a reference CRN which exhibits a desired continuous behaviour, the principles of Proportional-Integral-Derivative (PID) control are employed, in an attempt to control the behaviour of any arbitrary CRN. That is, given an input signal, represented by the concentration level of a species, a series of CRNs are constructed that in turn attempt to control the output concentration of an arbitrary CRN, such that the plant CRN exhibits the behaviour of the reference signal. A novel CRN implementation of a derivative component is provided, and proofs and simulation of its operation are given. This novel derivative component is used as a building block for a PID controller, enabling us to compare this with an existing PI controller, and show how negative feedback with a PID controller can be implemented in CRNs. The effectiveness of this architecture on a microRNA regulated gene expression example is demonstrated, where the time evolution of a protein is controlled by acting on the expression of mRNA and microRNA using a CRN reference signal provided by our synthesis method.

Study of Greek epic poetry in late antiquity

Plasma testosterone, luteinizing hormone (L.H.) and follicle stimulating (F.S.H.) were measured before operation in 39 males, aged 26 to 53 years with 2 to 5 children, seeking vasectomy for contraceptive purposes. These measurements were repeated at 3 and 12 months after vasectomy. Levels of L.H. at 12 months were significantly higher than pre-vasectomy levels, suggesting some degree of decreased Leydig cell function. However, testosterone and F.S.H. levels were similar at the three sampling times. Continuing yearly measurements for at least 5 years are planned.The long-term effect of vasectomy on testicular function was studied in 39 patients. At 12 months, luteinizing hormone levels were significantly higher than preoperative levels (p .005), which suggests some diminution of Leydig cell function. However, testosterone and follicle stimulating hormone levels did not vary markedly from preoperative levels. A 5-year program of yearly measurements of these hormones is planned.

Background: This study aimed to gain insight into the extent to which gaming is responsible for the underreporting of hospital-acquired infections (HAIs) in Mongolian hospitals, to identify gaming strategies used by health professionals, and to determine how gaming might be prevented. Methods: Eighty-seven health professionals, including policy- and hospital-level managers, doctors, nurses, and infection control practitioners, were recruited for 55 interviews and 4 group discussions in Mongolia in 2008. Results: All study participants were aware of gaming, which could occur via the following mechanisms: (1) doctors or nurses concealing HAI by overprescribing antibiotics or discharging patients early; (2) infection control practitioners failing to report HAI cases to hospital directors; and (3) hospital directors preventing reporting of HAI cases to the Ministry of Health. Gaming was consistently perceived to be a response to punitive performance evaluation by the Ministry of Health and penalization of hospitals and staff by the State Inspection Agency when HAIs were detected. Participants held divergent views regarding the best approach to reduce gaming, including excluding the current single indicator (ie, HAI rate) from the performance indicator list, developing multiple specific infection control indicators, improving the awareness of health managers regarding the causes of HAI, and increasing funding for infection control activities. Conclusion: Inclusion of the overall HAI rate in the targeted performance indicator set and the strict control and penalization of hospitals with reported HAI cases are factors that have contributed to gaming, which has resulted in deliberate, extreme underreporting of HAIs in Mongolian hospitals. Copyright © 2011 by the Association for Professionals in Infection. Control and Epidemiology, Inc. Published by Elsevier Inc. All rights reserved.

As with other areas of the public sector in Mongolia, the healthcare system has undergone significant structural and policy reforms since the early 1990s. The previous infection control system, characterised as a sanitary-epidemiological network, was dismantled with no replacement. A new infection control management system was established in 1997 with the adoption of infection control policies and guidelines, establishment of hospital infection control programmes in all major hospitals, training of health professionals and the commencement of passive surveillance of hospital-acquired infections (HAIs). Recent health statistics claim that HAIs occur in 0.01-0.02% of all hospital admissions with the highest percentage (0.05%) in tertiary hospitals in the capital city Ulaanbaatar, but this is very likely to be an underestimate. In 2002 the Government approved a national programme to establish a sentinel surveillance system for HAIs with improved laboratory-based monitoring. However, implementation has been delayed due to insufficient support from stakeholders and a shortage of resources and trained infection control professionals. Non-governmental infection control initiatives are limited by time and coverage. © 2010 The Hospital Infection Society.

Health statistics of Mongolia indicate that hospital-acquired infections (HAIs) occur in 0.01-0.05% of all hospital admissions. This is considerably lower than internationally reported rates. A one-day survey was conducted in two tertiary hospitals of Ulaanbaatar in September 2008 to estimate HAI prevalence, associated risk factors and patterns of antibiotic usage. Among 933 patients surveyed, 50 (5.4%) were diagnosed with HAI. Prevalence of surgical site infection was 1.1% (3.9% among surgical patients), bloodstream infection 0.3%, respiratory tract infection 1.3%, urinary tract infection 1.3%, and other HAI 1.4%. Microbiological investigations were only documented for 18.9% of all patients. A total of 558 patients (59.8%) were taking 902 courses of antibiotics; 92.1% of patients were prescribed antibiotics without a sensitivity test. Multiple logistic regression analysis revealed that HAI was significantly associated with the admission source, the hospital, length of hospital stay, surgical and other invasive procedures, urinary catheters and other indwelling devices. The study results were comparable with reports from some other developing countries and confirm that official statistics underestimate the true frequency of HAI in Mongolia. © 2010 The Hospital Infection Society.