Assessment of climate change statistical downscaling methods: Application and comparison of two statistical methods to a single site in Lisbon

Dissertação apresentada na Faculdade de Ciências e Tecnologia da Universidade Nova de Lisboa para a obtenção do grau de Mestre em Engenharia do Ambiente

Exploraty Multivariate Statistical Methods Applied to Pharmaceutical Industry CRM Data

Dissertação apresentada como requisito parcial para obtenção do grau de Mestre em Estatística e Gestão de Informação

Extremum estimators and stochastic optimization methods

Submitted in partial fulfillment for the Requirements for the Degree of PhD in Mathematics, in the Speciality of Statistics in the Faculdade de Ciências e Tecnologia

Multiple sequence alignment correction using constraints

Trabalho apresentado no âmbito do European Master in Computational Logics, como requisito parcial para obtenção do grau de Mestre em Computational Logics

Digital Divide Across the European Union

Dissertação apresentada como requisito parcial para obtenção do grau de Mestre em Estatística e Gestão de Informação

C-reactive protein in early risk assessment in acute pancreatitis

ABSTRACT:C-reactive protein (CRP) has been widely used in the early risk assessment of patients with acute pancreatitis (AP), but unclear aspects about its prognostic accuracy in this setting persist. This project evaluated first CRP prognostic accuracy for severity, pancreatic necrosis (PNec), and in-hospital mortality (IM) in AP in terms of the best timing for CRP measurement and the optimal CRP cutoff points. Secondly it was evaluated the CRP measured at approximately 24 hours after hospital admission (CRP24) prognostic accuracy for IM in AP individually and in a combined model with a recent developed tool for the early risk assessment of patients with AP, the Bedside Index for Severity in AP (BISAP). Two single-centre retrospective cohort studies were held. The first study included 379 patients and the second study included 134 patients. Statistical methods such as the Hosmer-Lemeshow goodness-of-fit test, the area under the receiver-operating characteristic curve, the net reclassification improvement, and the integrated discrimination improvement were used. It was found that CRP measured at approximately 48 hours after hospital admission (CRP48) had a prognostic accuracy for severity, PNec, and IM in AP better than CRP measured at any other timing. It was observed that the optimal CRP48 cutoff points for severity, PNec, and IM in AP varied from 170mg/l to 190mg/l, values greater than the one most often recommended in the literature – 150mg/l. It was found that CRP24 had a good prognostic accuracy for IM in AP and that the cutoff point of 60mg/l had a negative predictive value of 100%. Finally it was observed that the prognostic accuracy of a combined model including BISAP and CRP24 for IM in AP could perform better than the BISAP alone model. These results might have a direct impact on the early risk assessment of patients with AP in the daily clinical practice.--------- RESUMO: A proteina c-reactiva (CRP) tem sido largamente usada na avaliação precoce do risco em doentes com pancreatite aguda (AP), mas aspectos duvidosos acerca do seu valor prognóstico neste contexto persistem. Este projecto avaliou primeiro o valor prognóstico da CRP para a gravidade, a necrose pancreática (PNec) e a mortalidade intra-hospitalar (IM) na AP em termos do melhor momento para efectuar a sua medição e dos seus pontos-de-corte óptimos. Em segundo lugar foi avaliado o valor prognóstico da proteína c-reactiva medida aproximadamente às 24 horas após a admissão hospitalar (CRP24) para a IM na AP isoladamente e num modelo combinado, que incluiu uma ferramenta de avaliação precoce do risco em doentes com AP recentemente desenvolvida, o Bedside Index for Severity in Acute Pancreatitis (BISAP). Dois estudos unicêntricos de coorte retrospectivo foram realizados. O primeiro estudo incluiu 379 doentes e o segundo estudo incluiu 134 doentes. Metodologias estatísticas como o teste de Hosmer-Lemeshow goodness-of-fit, a area under the receiver-operating characteristic curve, o net reclassification improvement e o integrated discrimination improvement foram usadas. Verificou-se que a CRP medida às 48 horas após a admissão hospitalar (CRP48) teve um valor prognóstico para a gravidade, a PNec e a IM na AP melhor do que a CRP medida em qualquer outro momento. Observou-se que os pontos de corte óptimos da CRP48 para a gravidade, a PNec e a IM na AP variaram entre 170mg/l e 190mg/l, valores acima do valor mais frequentemente recomendado na literatura – 150mg/l. Verificou-se que a CRP medida aproximadamente às 24 horas após a admissão hospitalar (CRP24) teve um bom valor prognóstico para a IM na AP e que o ponto de corte 60mg/l teve um valor preditivo negativo de 100%. Finalmente observou-se que o valor prognóstico de um modelo combinado incluindo o BISAP e a CRP24 para a IM na AP pode ter um desempenho melhor do que o do BISAP isoladamente. Estes resultados podem ter um impacto directo na avaliação precoce do risco em doentes com AP na prática clínica diária.

Behavioural assessment impacts on the performance results

A Work Project, presented as part of the requirements for the Award of a Masters Degree in Management from the NOVA – School of Business and Economics

Digital Divide Across the European Union

Dissertação apresentada como requisito parcial para obtenção do grau de Mestre em Estatística e Gestão de Informação

Spatial Patterns and Irregularities of the electoral data: general elections in Canada

Dissertation submitted in partial fulfillment of the requirements for the Degree of Master of Science in Geospatial Technologies.

Politicization of non-political events

Spatial analysis and social network analysis typically take into consideration social processes in specific contexts of geographical or network space. The research in political science increasingly strives to model heterogeneity and spatial dependence. To better understand and geographically model the relationship between “non-political” events, streaming data from social networks, and political climate was the primary objective of the current study. Geographic information systems (GIS) are useful tools in the organization and analysis of streaming data from social networks. In this study, geographical and statistical analysis were combined in order to define the temporal and spatial nature of the data eminating from the popular social network Twitter during the 2014 FIFA World Cup. The study spans the entire globe because Twitter’s geotagging function, the fundamental data that makes this study possible, is not limited to a geographic area. By examining the public reactions to an inherenlty non-political event, this study serves to illuminate broader questions about social behavior and spatial dependence. From a practical perspective, the analyses demonstrate how the discussion of political topics fluсtuate according to football matches. Tableau and Rapidminer, in addition to a set basic statistical methods, were applied to find patterns in the social behavior in space and time in different geographic regions. It was found some insight into the relationship between an ostensibly non-political event – the World Cup - and public opinion transmitted by social media. The methodology could serve as a prototype for future studies and guide policy makers in governmental and non-governmental organizations in gauging the public opinion in certain geographic locations.

Changes in prescribing patterns of benzodiazepines after training of general practitioners

RESUMO: Introdução: As benzodiazepinas são os fármacos ansiolíticos e hipnóticos mais utilizados. O elevado consumo destes fármacos tem representado uma preocupação devido aos efeitos secundários do seu uso prolongado e dependência. Portugal tem a maior utilização de benzodiazepinas na Europa. Este estudo pretende analisar a alteração do padrão de prescrição de benzodiazepinas após uma intervenção com clínicos gerais. Métodos: A intervenção consistiu numa sessão educacional a um grupo de clínicos gerais. Foi comparado o padrão de prescrição de benzodiazepinas dos médicos intervencionados com o de um grupo de médicos não intervencionado da mesma região e com o de um grupo de médicos não intervencionados de outra região. Analisaram-­‐se as prescrições de 12 meses antes e depois da intervenção. A análise do padrão de prescrição utilizou como metodologia a Dose Diária Definida (DDD) e a Dose Diária Definida por 1000 pacientes por dia (DHD). A análise estatística recorreu a métodos de regressão segmentada. Resultados: Houve uma diminuição no padrão de prescrição de benzodiazepinas no grupo intervencionado após a intervenção (p=0.005). Houve também uma redução no padrão de prescrição no grupo não intervencionada da mesma região (p=0.037) e no grupo não-intervencionado da região diferente (p=0.010). Analisando por género, prescritores do género feminino prescrevem uma quantidade maior de benzodiazepinas. Os clínicos gerais do género feminino intervencionados tiveram a maior redução na prescrição após a intervenção (p=0.008). Discussão: Os dados demonstraram que a intervenção reduziu a prescrição de benzodiazepinas após a intervenção. A diminuição geral do padrão de prescrição poderá ser explicada pelo efeito de Hawthorne ou pela contaminação entre os três grupos de clínicos gerais. Os dados disponíveis não explicam as diferenças nos padrões de prescrição por género. Conclusão: Este estudo demonstra como uma única intervenção tem um impacto positivo na melhoria dos padrões de prescrição. A replicação desta intervenção poderá representar uma oportunidade para alterar a prescrição de benzodiazepinas em Portugal. -----------------------------ABSTRACT: Introduction: Benzodiazepines are the most utilized anxiolytic and hypnotic drugs. The high consumption of benzodiazepines has been a concern due to the reported side effects of long-­‐term use and dependence. Portugal has the highest benzodiazepine utilisation in Europe. This study aims to analyse the change in General Practitioners’ (GPs) benzodiazepine prescription pattern after na intervention period. Methods: An educational session was delivered to a group of intervened GPs. The benzodiazepine prescription pattern of the intervened group was compared to the pattern of a non-­‐intervened matched group from the same region, and to the pattern of another non-­‐intervened matched group from a diferente region. The research time frame was 12 month before and after intervention. The analysis of the prescription trends used the Defined Daily Dose (DDD) and Defined Daily Dose per 1000 patients per day (DHD) methodology. The statistical methods consisted of segmented regression analysis. Results: There was a decrease in benzodiazepine prescription pattern of intervened GPs after intervention (p=0.005). There was also a decrease in benzodiazepine prescription pattern for the non-­‐intervened group from the same region (p=0.037) and for the non-­‐ intervened group from a diferente region (p=0.010). Concerningthe analysis by gender, female gender prescribed a higher amount of benzodiazepines. The intervened female gender prescribers presented the highest decrease in prescription trend after intervention (p=0.008). Discussion: The data demonstrated that the intervention was effective in reducing benzodiazepine prescription after intervention. The general decrease in prescription trend might be explained by a Hawthorne effect or a contamination effect between the three groups of GPs. The available data couldn´t explain the diferences in prescription patterns by gender. Conclusion: This study demonstrates how a single intervention has a positive impact on improving prescription trends. The replication of this intervention might be an opportunity to changing the worrying benzodiazepine utilisation in Portugal.

Automatic learning for the classification of chemical reactions and in statistical thermodynamics

This Thesis describes the application of automatic learning methods for a) the classification of organic and metabolic reactions, and b) the mapping of Potential Energy Surfaces(PES). The classification of reactions was approached with two distinct methodologies: a representation of chemical reactions based on NMR data, and a representation of chemical reactions from the reaction equation based on the physico-chemical and topological features of chemical bonds. NMR-based classification of photochemical and enzymatic reactions. Photochemical and metabolic reactions were classified by Kohonen Self-Organizing Maps (Kohonen SOMs) and Random Forests (RFs) taking as input the difference between the 1H NMR spectra of the products and the reactants. The development of such a representation can be applied in automatic analysis of changes in the 1H NMR spectrum of a mixture and their interpretation in terms of the chemical reactions taking place. Examples of possible applications are the monitoring of reaction processes, evaluation of the stability of chemicals, or even the interpretation of metabonomic data. A Kohonen SOM trained with a data set of metabolic reactions catalysed by transferases was able to correctly classify 75% of an independent test set in terms of the EC number subclass. Random Forests improved the correct predictions to 79%. With photochemical reactions classified into 7 groups, an independent test set was classified with 86-93% accuracy. The data set of photochemical reactions was also used to simulate mixtures with two reactions occurring simultaneously. Kohonen SOMs and Feed-Forward Neural Networks (FFNNs) were trained to classify the reactions occurring in a mixture based on the 1H NMR spectra of the products and reactants. Kohonen SOMs allowed the correct assignment of 53-63% of the mixtures (in a test set). Counter-Propagation Neural Networks (CPNNs) gave origin to similar results. The use of supervised learning techniques allowed an improvement in the results. They were improved to 77% of correct assignments when an ensemble of ten FFNNs were used and to 80% when Random Forests were used. This study was performed with NMR data simulated from the molecular structure by the SPINUS program. In the design of one test set, simulated data was combined with experimental data. The results support the proposal of linking databases of chemical reactions to experimental or simulated NMR data for automatic classification of reactions and mixtures of reactions. Genome-scale classification of enzymatic reactions from their reaction equation. The MOLMAP descriptor relies on a Kohonen SOM that defines types of bonds on the basis of their physico-chemical and topological properties. The MOLMAP descriptor of a molecule represents the types of bonds available in that molecule. The MOLMAP descriptor of a reaction is defined as the difference between the MOLMAPs of the products and the reactants, and numerically encodes the pattern of bonds that are broken, changed, and made during a chemical reaction. The automatic perception of chemical similarities between metabolic reactions is required for a variety of applications ranging from the computer validation of classification systems, genome-scale reconstruction (or comparison) of metabolic pathways, to the classification of enzymatic mechanisms. Catalytic functions of proteins are generally described by the EC numbers that are simultaneously employed as identifiers of reactions, enzymes, and enzyme genes, thus linking metabolic and genomic information. Different methods should be available to automatically compare metabolic reactions and for the automatic assignment of EC numbers to reactions still not officially classified. In this study, the genome-scale data set of enzymatic reactions available in the KEGG database was encoded by the MOLMAP descriptors, and was submitted to Kohonen SOMs to compare the resulting map with the official EC number classification, to explore the possibility of predicting EC numbers from the reaction equation, and to assess the internal consistency of the EC classification at the class level. A general agreement with the EC classification was observed, i.e. a relationship between the similarity of MOLMAPs and the similarity of EC numbers. At the same time, MOLMAPs were able to discriminate between EC sub-subclasses. EC numbers could be assigned at the class, subclass, and sub-subclass levels with accuracies up to 92%, 80%, and 70% for independent test sets. The correspondence between chemical similarity of metabolic reactions and their MOLMAP descriptors was applied to the identification of a number of reactions mapped into the same neuron but belonging to different EC classes, which demonstrated the ability of the MOLMAP/SOM approach to verify the internal consistency of classifications in databases of metabolic reactions. RFs were also used to assign the four levels of the EC hierarchy from the reaction equation. EC numbers were correctly assigned in 95%, 90%, 85% and 86% of the cases (for independent test sets) at the class, subclass, sub-subclass and full EC number level,respectively. Experiments for the classification of reactions from the main reactants and products were performed with RFs - EC numbers were assigned at the class, subclass and sub-subclass level with accuracies of 78%, 74% and 63%, respectively. In the course of the experiments with metabolic reactions we suggested that the MOLMAP / SOM concept could be extended to the representation of other levels of metabolic information such as metabolic pathways. Following the MOLMAP idea, the pattern of neurons activated by the reactions of a metabolic pathway is a representation of the reactions involved in that pathway - a descriptor of the metabolic pathway. This reasoning enabled the comparison of different pathways, the automatic classification of pathways, and a classification of organisms based on their biochemical machinery. The three levels of classification (from bonds to metabolic pathways) allowed to map and perceive chemical similarities between metabolic pathways even for pathways of different types of metabolism and pathways that do not share similarities in terms of EC numbers. Mapping of PES by neural networks (NNs). In a first series of experiments, ensembles of Feed-Forward NNs (EnsFFNNs) and Associative Neural Networks (ASNNs) were trained to reproduce PES represented by the Lennard-Jones (LJ) analytical potential function. The accuracy of the method was assessed by comparing the results of molecular dynamics simulations (thermal, structural, and dynamic properties) obtained from the NNs-PES and from the LJ function. The results indicated that for LJ-type potentials, NNs can be trained to generate accurate PES to be used in molecular simulations. EnsFFNNs and ASNNs gave better results than single FFNNs. A remarkable ability of the NNs models to interpolate between distant curves and accurately reproduce potentials to be used in molecular simulations is shown. The purpose of the first study was to systematically analyse the accuracy of different NNs. Our main motivation, however, is reflected in the next study: the mapping of multidimensional PES by NNs to simulate, by Molecular Dynamics or Monte Carlo, the adsorption and self-assembly of solvated organic molecules on noble-metal electrodes. Indeed, for such complex and heterogeneous systems the development of suitable analytical functions that fit quantum mechanical interaction energies is a non-trivial or even impossible task. The data consisted of energy values, from Density Functional Theory (DFT) calculations, at different distances, for several molecular orientations and three electrode adsorption sites. The results indicate that NNs require a data set large enough to cover well the diversity of possible interaction sites, distances, and orientations. NNs trained with such data sets can perform equally well or even better than analytical functions. Therefore, they can be used in molecular simulations, particularly for the ethanol/Au (111) interface which is the case studied in the present Thesis. Once properly trained, the networks are able to produce, as output, any required number of energy points for accurate interpolations.

Genetics of berry colour and anthocyanin content variation in grapevine (Vitis vinifera L.subsp. vinifera)

Dissertation presented in fulfillment of the requirements for the Degree of Doctor of Philosophy in Biology (Molecular Genetics) at the Instituto de Tecnologia Química e Biológica da Universidade Nova de Lisboa

Numerical methods and tangible interfaces for pollutant dispersion simulation

Dissertação apresentada para obtenção do Grau de Doutor em Engenharia do Ambiente, pela Universidade Nova de Lisboa, Faculdade de Ciências e Tecnologia