Abstract: Campylobacter infections are the main bacterial cause of gastroenteritis in the UK, causing an estimated 500 thousand cases per year. Health authorities investigate outbreaks to identify the source, control the spread and understand the cause. Outbreak detection mechanisms are potentially improved by the increasing availability of whole-genome sequence alongside other epidemiological data. However, techniques mixing genomics and other epidemiological factors are still underdeveloped. This project aims to develop and apply outbreak detection methods using surveillance data collected from two regions in the UK. The approaches proposed in this thesis are based on an existing spatialtemporal Bayesian hierarchical model, where cases are labelled as potential outbreaks if they comprise an elevated number of cases compared to the expected sporadic count. The model is adjusted to include genetic data using Gaussian random fields, exploiting the capacity of whole-genome sequencing to discriminate closely related isolates. Moreover, a Markov Chain Monte Carlo algorithm is implemented to obtain the posterior distribution of the model parameters. In particular, a sampling strategy is proposed to improve the convergence of the chain for the parameters describing the Gaussian random field. The project dataset is analysed using a spatial-temporal, a spatial-genetic and a temporalgenetic version of the model, where each version explores different types of outbreaks. The proposed approach demonstrates how to organise genetic sequences into a highdimensional structure and incorporate them into a Bayesian framework. Also, the MCMC sampling algorithm improves the mixing of the chain to estimate the posterior distribution of the model parameters. Finally, all model versions provide the probability that each reported infection is part of a potential outbreak. Comparing the potential outbreaks found by each model provides insights to estimate the real outbreaks. It also identifies cases that are potentially part of a diffuse real outbreak hard to detect by existing approaches. Despite the capability of the model, it requires predefined outbreak sizes and therefore is not flexible at capturing many shapes. Autocorrelated models are a potential improvement to be explored. Resumen: Las infecciones por Campylobacter son la principal causa bacteriana de gastroenteritis en el Reino Unido, provocando un estimado de 500 mil casos por año. Las autoridades sanitarias investigan los brotes para identificar la fuente, controlar la propagación y comprender la causa. Los mecanismos de detección de brotes se mejoran potencialmente con la creciente disponibilidad de datos del genoma completo de las bacterias junto con otros datos epidemiológicos. Sin embargo, las técnicas que combinan genética y otros factores epidemiológicos aún están poco desarrolladas. Este proyecto tiene como objetivo desarrollar y aplicar métodos de detección de brotes utilizando datos de vigilancia recopilados en dos regiones del Reino Unido. Los enfoques propuestos en esta tesis se basan en un modelo jerárquico bayesiano espacio-temporal existente, donde los casos se etiquetan como brotes potenciales si presentan un número elevado de casos mayor al recuento esporádico esperado. El modelo se ajusta para incluir datos genéticos utilizando campos aleatorios Gaussianos, aprovechando la capacidad del genoma completo para discriminar casos estrechamente relacionados. Además, se implementa usando algoritmos de Markov Chain Monte Carlo para obtener la distribución posterior de los parámetros del modelo. En particular, se propone una estrategia de muestreo para mejorar la convergencia de la cadena de Markov para los parámetros del campo Gaussiano. El conjunto de datos del proyecto se analiza utilizando una versión espacial temporal, una espacial-genética y una temporal-genética del modelo, donde cada versión explora diferentes tipos de brotes. El enfoque propuesto demuestra cómo organizar secuencias genéticas en una estructura de múltiples dimensiones e incorporarlas en un marco bayesiano. Además, el algoritmo de muestreo MCMC mejora la mezcla de la cadena para estimar la distribución posterior de los parámetros del modelo. Por último, todas las versiones del modelo estiman la probabilidad de que cada infección sea parte de un posible brote. La comparación entre los posibles brotes encontrados por cada modelo proporciona información para estimar cuáles son los brotes reales. También identifica casos que son potencialmente parte de un brote real difícil de detectar por los enfoques existentes. A pesar de las múltiples ventajas del modelo, este requiere tamaños de brotes predefinidos y, por lo tanto, no es flexible para capturar brotes de formas irregulares. Los modelos autocorrelacionados proveen una mejora potencial para el modelo que debe explorarse.

Abstract: The growing demand for sustainable load-bearing materials drives the need for understanding the various design considerations these pose within the modern built environment. Engineered bamboo is a material with outstanding physical and mechanical properties, in addition to producing a minimum carbon footprint. However, extensive research is needed before engineered bamboo can be used with the confidence conferred to other more conventional building construction materials. When aiming for higher and larger bamboo-based structures, loadbearing behaviour during and after fire becomes a key consideration. This paper describes the outcomes of a comprehensive study conducted to understand the mechanical behaviour of bamboo (Phyllostachys pubescens species) at elevated temperatures; more specifically investigating the reduction of compressive and tensile strength, as well as the Modulus of Elasticity (MoE) up to 250 °C. Findings from this work show that at 200 °C, bamboo retains 20%, 42% and 70% of the compressive strength, tensile strength and modulus of elasticity at ambient, respectively. The results presented herein, which provide thorough understanding of strength and elasticity reduction at elevated temperatures, enable the development of stress-strain constitutive models that will constitute the basis for designing fire-safe bamboo structures.

Abstract: The remarkable high strength-to-weight ratio of bamboo and the recent innovations and development of laminated bamboo products has enabled the potential for new larger and taller structural bamboo applications. However, there is little understanding of the behaviour and performance of bamboo structures during and after a fire. To date, bamboo has been used widely in low-rise constructions where structural integrity during and after a fire does not pose a risk to the life safety of occupants, fire service intervention, and significant impact on property protection or business continuity. Nevertheless, a comprehensive understanding of the structural fire performance is crucial if bamboo is to be widely used in the construction of mid- and high-rise load-bearing structures. An extensive literature review on the fire performance of bamboo structures shown herein identifies that novel fire testing methods were needed to investigate the mechanical response of bamboo, primarily to study the reduction in the strength and stiffness at elevated temperatures. Past research outcomes have evidenced the need for constitutive stress-strain models to describe the mechanical behaviour of bamboo at elevated temperatures, as well as numerical models to predict the response of the bamboo loadbearing members during and after a fire. Several small and bench-scale experimental studies were conducted to understand the reduction in the mechanical properties of bamboo at elevated temperatures. Compressive tests were performed for round and laminated bamboo for test samples at temperatures between ambient and 250ºC. Round bamboo strips were tested in tension to understand the reduction in the tensile strength at elevated temperatures. A dynamic mechanical analysis of bamboo strips was carried out to investigate the decrease in the modulus of elasticity at elevated temperatures, and to compare this against the results of bench-scale tests to analyse the effect of a different test set-up and heat transfer scenario. With the experimental results of the reduction of the compressive and tensile strength of bamboo, as well as the modulus of elasticity, constitutive stress-strain models were proposed to correlate the stress-and strain relationship at elevated temperatures. Equations for the reduction factors of strength and modulus of elasticity were presented, as well as stress-strain curves. The linear elastic behaviour of bamboo under tension and compression, and the elastoplastic behaviour under compression were analysed to quantify the effects of elevated temperatures. Another series of experiments were conducted to investigate the performance of laminated bamboo beams exposed to fire. The bamboo beams were loaded at different stress-levels whilst exposed to a constant indecent heat flux. The beams were fully instrumented to measure the loss of cross-section, the strain profiles, and the deflections at the mid-span during the heating exposure. Finally, a numerical model combining the thermal performance, and the mechanical behaviour of bamboo at elevated temperatures was developed to predict the stress-strain profiles, the deflection, and the bending capacity of bamboo beams during a fire. The results of the numerical model were compared against the results obtained from the experimental tests, and the model showed a good agreement with the experimental results. The proposed framework shows that structural analysis can be carried out if heat transfer analysis and constitutive stress-strain models are available to predict the thermomechanical response of laminated bamboo beams. Findings from this work provide a good understanding of the reduction on the strength and elasticity of bamboo at elevated temperatures. These results will enable predictions to be made for the bending capacity of bamboo beams during a fire, and they will constitute the basis for designing fire-safe bamboo structures.

Resumen: Esta investigación analiza las principales razones que han dado lugar a la prohibición del uso velo islámico en el espacio público francés, cuestionando su finalidad y la proporción que guarda en relación a los derechos de libertad de expresión, libertad de culto, y libertad ideológica. La discusión constantemente se encuentra en la búsqueda por establecer cuando en ejercicio de una garantía propia se puede violentar una libertad ajena, por lo cual, la relación entre derechos y limites debe ser cuidadosamente observada.

Abstract: The school buildings in Colombia are built based on geographical locations and regional construction systems. However, external weather conditions and building design can have a significant impact on the thermal comfort of students, which affects the academic performance and productivity. This paper investigates the thermal comfort performance for an educational building in a hot and humid city in Cucuta, Colombia, built under national guidelines. This school is a concrete structure without mechanical cooling. However, field observation discovered that 82% of the time students experienced thermal discomfort. To investigate causes and provide mitigation strategies, a whole building energy simulation is conducted. Design Builder is used to evaluate the indoor thermal conditions compared to outdoor data collected. ASHRAE 55 adaptative model is used for the evaluation. It is found that 79% of the time the thermal conditions are outside the acceptable range during the year. The effect of mitigation measures i.e., occupancy, roof insulation, and natural ventilation rates are investigated through simulations. It is found that occupancy and natural ventilation rate have a significant impact on the indoor temperature and relative humidity, and thus the thermal comfort. Passive design strategies are proposed in optimizing the school building design to meet ASHARE-55 requirements.

Abstract: Background: Seasonal Influenza is considered to be a cyclic and ordered sequence of values, influenced by external factors that can be predicted and used to detect disease outbreaks and monitoring. In machine learning, the key challenges that limit these analyses are in model explainability and limitations associated with ecological bias. Aim: Determine the best environmental variable selection method to predict Seasonal Influenza in Norway, using an environmental medicine approach combined with machine learning techniques. Methods: This is a quasi-experimental study that compares three approaches (non-variable selection, isolate component, and multipollutant mixture), represented in five methods (univariable, bivariable, multivariable AME, multivariable PCA, multivariable LDA). Per method, the best co-variable combination will be performed, following the internal rules of each method. The best covariable combination is the result of three components: variable selection, validation data set and lag. The first one involves 13 environmental variables (temperature, relative humidity, specific humidity, air pressure, wind speed, precipitation, CO, NO, NO2, O3, PM10, PM2,5 and SO2); second one compares a test dataset compiled from the data from 2019, last year (2018) and a synthetic environmental (avg. 2013-2018) data set in the validation process; and the third one compares a combination of lag from 0 to 12. All the predictions are made using ARIMA algorithm. The evaluation is given in terms of MAE, MSE, RMSE, OR. The training set is from 2 Jun 2013 (week 22/2013) to 28 May 2018 (week 21/2018), and test set is from (week 22/2018) to (week 21/2019) with a window of predictions of 52 weeks. Results: The increment of dimensionality in the environmental variable selection introduce different noise levels and optimize the prediction. Considerations that impact the explainability, usability, ecological bias and performance will be described. Conclusion: The increment of dimensionality in the variable selection has a better impact on performance than using complex algorithms.

Abstract: Objective: To analyze the population prevalence and birth prevalence of oral clefts in Colombia from 2009 to 2017. Methods: A cross-sectional study using information from the National Administrative Records of Colombia. The data came from 2 types of administrative records (Surveillance System and the Individual Registry of Service Provision) and the oral health national survey. Population prevalence and birth prevalence by type of cleft lip and/or cleft (CL/P) ratios were calculated using Poisson distribution for count data and to assess stationary tests on time series (Dickey-Fuller) and (Phillips-Perron) was used. Results: Population prevalence in Colombia was 3.27 per 10 000 inhabitants (95% confidence interval [CI], 3.21-3.32) and birth prevalence was 6.0 per 10 000 live births (95% CI, 5.67-6.35). Bogot´a have the highest population prevalence with CL/P. In the analysis of trends for the prevalence proportion by type of clefts in newborn babies with cleft, it was observed that the highest proportion was for babies with CLP. Cleft lip (CL) has increased from 17.4% in 2014 to 34.2% in 2017, cleft palate (CP) has decreased from 32.9% to 20.2%; and CLP changed from 49.6% to 45.5% in the same period. Conclusions: The population prevalence was 3.27 per 10 000 inhabitants. Births prevalence was 6.0 per 10 000 live births, and Orinoquia and Amazonia have higher rates than the national average. The administrative registers are adequate systems to know the behavior of oral clefts. The CL/P had a nonstationary trend during the period 2014 to 2017.

Abstract: Objective: Determine the population prevalence and trends of cleft lip and / or palate (CL/P) by department for Colombia in the period 2009 - 2015. Methods: Prevalence study based on Individual Registry of Health Services in general population from 2009 to 2015. All people diagnosed with CL/P were included for all ages, type of diagnosis and any type of health services in the mentioned period. The prevalence rate was calculated by period and point for each year, for each department and according to the type of cleft. Stationarity on time series was evaluated using (Dickey-Fuller) and (Phillips-Perron), the trends and prevalence ratios were calculated using Poisson regression. Results: 15,225 people with CL/P were identified, where 53.3% were men. The national period prevalence of CL/P is 3.37 per 10 000 (IC95%: 3.3-3.4) with upward trend (PR = 1.34 95% CI: 1.0 - 1.8 p = 0.05) and non-stationary behavior. The national period prevalence of CL is 0.93 per 10 000, CP 1.17 per 10 000 and CLP 1.26 per 10 000, where CLP is sub classify into CLPu (0.83 per 10 000), and CLPb (0.43 per 10 000). At the departmental level, the highest CL/P prevalence is Guaviare (11.2 95% CI: 8.6 – 14.2), followed by Guainía (8.4 95% CI:5.4 – 12.2) and the lowest Quindío (0.49 95% CI: 0.3 – 0.7) Conclusions: In Colombia, the national period prevalence of CL/P is 3.37 per 10 000 with upward trend at national level indicates an increase on prevalence from 2009 to 2015.

Abstract: Seasonal influenza is an infectious disease of multi-causal etiology and a major cause of mortality worldwide that has been associated with environmental factors. In the attempt to model and predict future outbreaks of seasonal influenza with multiple environmental factors, we face the challenge of increased dimensionality that makes the models more complex and unstable. In this paper, we propose a nowcasting and forecasting framework that compares the theoretical approaches of Single Environmental Factor and Multiple Environmental Factors. We introduce seven solutions to minimize the weaknesses associated with the increased dimensionality when predicting seasonal influenza activity levels using multiple environmental factors as external proxies. Our work provides evidence that using dimensionality reduction techniques as a strategy to combine multiple datasets improves seasonal influenza forecasting without the penalization of increased dimensionality.

Abstract: Introgression has been proposed as an essential source of adaptive genetic variation. However, a key barrier to adaptive introgression is that recombination can break down combinations of alleles that underpin many traits. This barrier might be overcome in supergene regions, where suppressed recombination leads to joint inheritance across many loci. Here, we study the evolution of a large supergene region that determines a major social and ecological trait in Solenopsis fire ants: whether colonies have one queen or multiple queens. Using coalescent-based phylogenies built from the genomes of 365 haploid fire ant males, we show that the supergene variant responsible for multiple-queen colonies evolved in one species and repeatedly spread to other species through introgressive hybridization. This finding highlights how supergene architecture can enable a complex adaptive phenotype to recurrently permeate species boundaries.