No description available.

Genome-wide association studies have been used extensively to study hundreds of phenotypes and have determined thousands of associated SNPs whose underlying biology and causation is as yet largely unknown. Many previous studies attempted to clarify the causal biology by investigating overlaps of trait-associated variants with functional annotations, but lacked statistical rigor and examined incomplete subsets of available functional annotations. Additionally, it has been difficult to disentangle the relative contributions of different annotations that may show strong correlations with one another. In this thesis, we address these shortcomings and strengthen and extend the obtained results. Two methods, permutations and logistic regression, are applied in statistically rigorous analyses of genomic annotations and their observed enrichment or depletion of trait-associated SNPs. The genomic annotations range from genic regions and regulatory features to measures of conservation and aspects of chromatin structure. Logistic regressions in a number of trait-specific subsets identify genomic annotations influencing SNPs associated with both normal variation (e.g., eye or hair colour) and diseases, suggesting some generalities in the biological underpinnings of phenotypes. SNPs associated with phenotypes of the immune system are investigated and the results highlight the distinct aetiology for this subset. Despite the heterogeneity of the studied cancers, SNPs associated to different cancers are particularly enriched for conserved regions, unlike all other trait-subsets. Nonetheless, chromatin states are, perhaps surprisingly, among the most influential genomic annotations in all trait-subsets. Evolutionary conserved regions are rarely within the top genomic annotations despite their widespread use in prioritisation methods for follow-up studies. We identify a common set of enriched or depleted genomic annotations that significantly influence all traits, but also highlight trait-­‐specific differences. These annotations may be used for the computational prioritisation of variants implicated in phenotypes of interest. The approaches developed for this thesis are further applied to studies of a specific human complex trait (height) and gene expression in atherosclerosis.

published_or_final_version / Microbiology / Doctoral / Doctor of Philosophy

Background: Healthcare-associated pneumonia (HCAP) guidelines were first proposed in 2005 but have not yet been validated. The objective of this study was to compare 30-day mortality and length of stay (LOS) in HCAP patients treated with either guideline-concordant HCAP (GC-HCAP) therapy or guideline-concordant community-acquired pneumonia (GC-CAP) therapy. Methods: We performed a retrospective cohort study of >150 hospitals in the Veterans Health Administration. Patients were included if they had ≥1 HCAP risk factor and received antibiotic therapy within 48 hours of admission. Patients were excluded if they received ICU care, had cardiovascular or respiratory organ failure, or received invasive mechanical ventilation and/or vasopressors. We determined independent risk factors for 30-day mortality with a multivariable logistic regression model including baseline characteristics, individual HCAP risk factors, comorbidities, and organ failure as dichotomous covariates. Propensity scores were calculated for the probability to receive GC-HCAP therapy and incorporated into a second logistic regression model. Results: A total of 15,071 patients met study criteria and received GC-HCAP therapy (8.0%), GC-CAP therapy (75.7%), or non-guideline-concordant therapy (16.3%). GC-HCAP patients were more likely to have neoplastic disease; whereas, GC-CAP patients had a higher prevalence of other comorbidities, tobacco use, and recent medication use. In multivariable regression, recent hospital admission (OR 2.47, 95% CI 2.10-2.91) and GC-HCAP therapy (2.13, 1.82-2.48) were the strongest predictors of 30-day mortality. Hematologic organ failure, non-invasive mechanical ventilation, neoplastic disease, renal organ failure, and cerebrovascular disease were also independent risk factors. Use of cardiovascular medications, inhaled corticosteroids, and tobacco were protective. GC-HCAP therapy continued to be an independent risk factor for 30-day mortality (OR 2.12, 95% CI 1.82-2.48) in the propensity score analysis. Conclusions: GC-HCAP therapy is not associated with improved survival in HCAP patients. / text

Clostridium difficile infection (CDI) is the leading cause of bacterial infectious diarrhea in nosocomial settings and approximately 25% of patients with CDI experience disease recurrence. Prior CDI epidemiological investigations are limited though. They do not reflect the burden of CDI in federal facilities, nor do they capture recent estimates on patient health outcomes. Furthermore, few studies have integrated CDI recurrence risk factors into a tool that clinicians can use to identify patients at risk for CDI recurrence. This study 1) described the epidemiology of CDI in the national Veterans Health Administration (VHA), 2) derived and validated a clinical prediction rule for 60-day CDI first recurrence, and 3) derived and validated a clinical prediction rule for 60-day CDI second recurrence. This was a retrospective cohort study of VHA beneficiaries with CDI between October 1, 2001 and September 30, 2012. VHA clinical and pharmacy data were integrated to develop several independent variables, including patient baseline demographics, CDI characteristics, comorbidities, concomitant medications, prior medications, prior hospitalization, hospital length of stay (LOS), and CDI severity. The dependent variables included 30/60/90-day mortality, and 30/60/90-day CDI recurrence. CDI incidence and outcomes were presented descriptively and compared using generalized linear regression models. CDI recurrence prediction rules were derived using multivariable logistic regression models and validated using the area under the receiver-operating-characteristic curve (AUROC), sensitivity, specificity, positive predictive value, and negative predictive value. Our study demonstrated that CDI first episodes, recurrences, and severity increased over the study period, while mortality decreased. Our CDI first recurrence prediction rule included the following predictor variables: dyslipidemia, CDI type, renal disease, hospital LOS <7 days, principal CDI diagnosis, concomitant gastric acid suppressors, and concomitant antibiotics. This model demonstrated moderate 60-day first recurrence discrimination (AUROC=0.62). Our CDI second recurrence prediction rule was similar in predictor variables and validity. In conclusion, CDI is an important, rapidly-emerging public health problem in the VHA. A clinical prediction rule might aid clinicians in directing preventative therapies to patients at high risk for CDI recurrence. / text

No description available.

No description available.

Multiple DNA transactions are at the center of almost all processes regulating the AAV life cycle. A common feature shared by all transactions is the binding of the large AAV Rep proteins Rep78/Rep68 onto DNA sites harboring multiple GCTC repeats. AAV mediated site-specific integration is contingent upon the formation of a productive complex between Rep78/Rep68 and the AAVS1 site located at chromosome 19. In order to understand the mechanistic details of the initial assembly process we carried out equilibrium binding experiments of Rep68 and its individual domains with a 42-mer AAVS1 site. Results show that although Rep68 binds AAVS1 with high affinity (69 nM), both the OBD and helicase individual domains bind DNA weakly with affinities of >>60μM and 22μM respectively under our experimental conditions. Mutant Rep68 proteins that have a defective oligomerization interface bind DNA poorly suggesting that productive binding requires both the concerted interaction of the individual domains with DNA and oligomerization. Moreover, we show that a minimal number of two repeats is required to form a stable complex. In addition, initial studies were done to characterize the interaction between Rep68 and the viral ITR DNA sequences using AUC and electron microscopy.

The communities of microorganisms associated with humans and other animals are characterized by a large degree of diversity and unexplained variation across individual hosts. While efforts to explain this variation in host-associated systems have focused heavily on the effects of host selection, community assembly theory emphasizes the role of dispersal and stochastic demographic processes, otherwise known as ecological drift. In this dissertation, I characterize the communities of microorganisms associated with the zebrafish, Danio rerio, intestine, and assess the importance of microbial dispersal and drift to their assembly. First, I describe changes in the composition and diversity of the zebrafish intestinal microbiome over zebrafish development and show that while host development is a major driver of community composition over time, there remains a large amount of unexplained variation among similar hosts of the same age. I go on to show that random dispersal and ecological drift alone in the absence of host selection are sufficient to explain a substantial amount of this variation, but the ability of these processes to predict the distribution of microorganisms across hosts decreases over host development. Finally, I present an experimental test of dispersal in host-associated systems, and show that not only does dispersal among individual zebrafish hosts have a large impact on the composition and diversity of associated microbial communities, but it can also overwhelm the effects of important host factors, such as the innate immune system. As a whole, this work demonstrates that the composition and diversity of microbial communities associated with animal hosts are not solely the result of selection by the host environment, but rather dispersal and stochastic processes have important and often overwhelming effects on their assembly. To fully understand the assembly of host-microbe systems, we must broaden our focus to include scales beyond that of an individual host and their associated microorganisms. / 10000-01-01

Ventilator-associated pneumonia (VAP) is a serious complication in critically ill patients; it can prolong intubation, increase intensive care unit and hospital length of stay, and increase mortality to twice the level of patients who do not develop VAP. The purpose of this project was to determine the effect of an evidence-based educational program to prevent VAP on ICU nurses' actual and documented practices for preventing VAP. The research questions addressed whether an educational program focused on VAP prevention will affect critical care nurses' compliance with a VAP prevention bundle, and whether the education will result in maintenance of a rate of zero cases of VAP per 1000 ventilator days. Data will be collected from all ICU patients intubated more than 24 hours and will include: (a) the frequency of oral care, (b) head-of-bed elevation of 30-45 degrees, (c) daily sedation vacation, (d) assessment of readiness for extubation, and (e) whether prophylaxis for deep vein thrombosis and for peptic ulcer disease was ordered. Observations of care will verify the accuracy of nurses' documentation in the medical record. A survey will assess nurse satisfaction with the educational program. Paired t tests will be used to compare the compliance of the nurses with each element of oral care and hygiene practices before and after the intervention. Analysis of variance will be calculated on the mean duration of ventilation, mean ICU and hospital length of stay, mortality before discharge, patient acuity, and rates of VAP per 1000 ventilator days. The goal of this project is a compliance rate of 90% or greater with the elements of the VAP prevention bundle, leading to decreased ventilator and ICU days, decreased morbidity, decreased mortality, and lower emotional distress. Positive social change will be accomplished through an immediate improvement in the lives of VAP-prone individuals.