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.

Prognostic modelling of breast cancer patients: a benchmark of predictive models with external validation

Dissertação apresentada para obtenção do Grau de Doutor em Engenharia Electrotécnica e de Computadores – Sistemas Digitais e Percepcionais pela Universidade Nova de Lisboa, Faculdade de Ciências e Tecnologia

New analytical methodologies for doping control – detection of anabolic androgenic steroids in human urine

Dissertation submitted to Faculdade de Ciências e Tecnologia - Universidade Nova de Lisboa in fulfilment of the requirements for the degree of Doctor of Philosophy (Biochemistry - Biotechnology)

Archaeological roman glasses: comparative characterisation by non-destructive analytical techniques

Dissertation presented in fulfilment of the requirements for the Master’s degree in Conservation and Restoration

Development and validation of a risk model for prediction of hazardous alcohol consumption in general practice attendees : the PredictAL study

Background: Little is known about the risk of progression to hazardous alcohol use in people currently drinking at safe limits. We aimed to develop a prediction model (predictAL) for the development of hazardous drinking in safe drinkers. Methods: A prospective cohort study of adult general practice attendees in six European countries and Chile followed up over 6 months. We recruited 10,045 attendees between April 2003 to February 2005. 6193 European and 2462 Chilean attendees recorded AUDIT scores below 8 in men and 5 in women at recruitment and were used in modelling risk. 38 risk factors were measured to construct a risk model for the development of hazardous drinking using stepwise logistic regression. The model was corrected for over fitting and tested in an external population. The main outcome was hazardous drinking defined by an AUDIT score >= 8 in men and >= 5 in women. Results: 69.0% of attendees were recruited, of whom 89.5% participated again after six months. The risk factors in the final predictAL model were sex, age, country, baseline AUDIT score, panic syndrome and lifetime alcohol problem. The predictAL model's average c-index across all six European countries was 0.839 (95% CI 0.805, 0.873). The Hedge's g effect size for the difference in log odds of predicted probability between safe drinkers in Europe who subsequently developed hazardous alcohol use and those who did not was 1.38 (95% CI 1.25, 1.51). External validation of the algorithm in Chilean safe drinkers resulted in a c-index of 0.781 (95% CI 0.717, 0.846) and Hedge's g of 0.68 (95% CI 0.57, 0.78). Conclusions: The predictAL risk model for development of hazardous consumption in safe drinkers compares favourably with risk algorithms for disorders in other medical settings and can be a useful first step in prevention of alcohol misuse.

A study on reducing digital piracy: an analytical network process approach with benefits, opportunities, costs and risks analysis

Dissertação para obtenção do Grau de Mestre em Engenharia e Gestão Industrial

Validation of NASA rotor 67 with openFOAM's transonic density-based solver

Dissertação para obtenção do Grau de Mestre em Engenharia Mecânica

Validation of alternative additives for paste production

Dissertation presented to confer Master Degree in Chemical and Biochemical Engineering

Validação e aplicação clínica de uma escala de cognição social na esquizofrenia = Validation and clinical use of a measure of social cognition in schizophrenia

RESUMO: A cognição social encontra-se frequentemente alterada na esquizofrenia. Esta alteração relaciona-se com a diminuição do funcionamento social,caracterizando-se quer por défices quer por vieses cognitivos sociais. No entanto, existem poucos instrumentos fiáveis e válidos para avaliar a cognição social na esquizofrenia, nomeadamente capazes de medir os vieses cognitivos sociais e a cognição social auto-relevante. Adicionalmente, as bases biológicas da disfunção social não estão totalmente esclarecidas. Evidências recentes sugerem que o peptídeo oxitocina (OXT) influencia o funcionamento social, e que esta relação poderá ser mediada pela cognição social. Este Trabalho de Projecto descreve a contribuição do autor para o desenvolvimento e avaliação psicométrica inicial de um novo instrumento de avaliação da cognição social, e a utilidade desta escala na investigação das associações entre a OXT e a capacidade e vieses cognitivos sociais. A Waiting Room Task (WRT), uma escala constituída por 26 vídeos sequenciais que simulam a experiência de observar outra pessoa numa sala de espera, foi administrada num estudo transversal com 61 doentes com esquizofrenia e 20 controlos saudáveis. Observou-se uma menor capacidade cognitiva social e um aumento dos vieses cognitivos sociais nos doentes com esquizofrenia, comparativamente aos controlos. Nos controlos e doentes com delírios, o desempenho na WRT correlacionou-se significativamente com os níveis de OXT. Esta correlação não se observou nos doentes sem delírios, sugerindo que o papel da OXT na cognição social poderá encontrar-se atenuado neste grupo. Estes achados fornecem suporte inicial para a adequação da WRT como instrumento de avaliação da cognição social na esquizofrenia, podendo ainda ser útil na investigação da sua base biológica. ------------ ABSTRACT: Social cognition is often impaired in schizophrenia. This impairment is related to poor social functioning and is characterized by both social cognitive deficits and biases. However, there are few reliable and valid measures of social cognition in schizophrenia, particularly measures of social cognitive bias and of self-relevant social cognition. Also, the biological bases of social dysfunction are not well understood. Emerging evidence suggests that the peptide oxytocin (OXT) influences social functioning, and that this relationship may be mediated by social cognition. This Research Project describes the author’s contribution to the development and initial psychometric testing of a new measure of social cognition, and the utility of this instrument to examine associations between OXT and social cognitive capacity and bias. The Waiting Room Task WRT), a video-based test comprising 26 sequential videos simulating the experience of facing another person in a waiting room, was administered in a cross-sectional study involving 61 patients with schizophrenia and 20 healthy controls. Social cognitive capacity was lower and social cognitive bias was increased in patients with schizophrenia compared with controls. Among controls and patients with delusions, performance on the WRT was significantly correlated with OXT level. This correlation was not found in patients without delusions suggesting that OXT’s role in social cognition may be blunted in this group. These findings provide initial support for the adequacy of the WRT as a measure for assessing social cognition in schizophrenia that may also be useful in understanding its biological underpinnings.

Exploring the experience of stigma in severe mental illness : a Portuguese contribution to the validation of a psychometric instrument

RESUMO: Fizemos uma análise da evolução do conceito de estigma, das suas correlações e das suas consequências e analisámos os instrumentos psicométricos utilizados para estudar experiências pessoais de estigma. Revimos os principais estudos de investigação sobre estigma em Portugal. Revimos, igualmente, os estudos relevantes utilizando o “Consumer Experiences of Stigma Questionnaire” (CESQ) e as propriedades psicométricas já documentadas. O nosso estudo teve como objetivos: explorar as experiências de estigma numa amostra portuguesa de pessoas com perturbação mental grave e contribuir para a documentação das propriedades psicométricas do “Consumer Experiences of Stigma Questionnaire” e para a validação da sua versão portuguesa. Fizemos um estudo transversal, descritivo e analítico, recolhemos dados sociodemográficos e clínicos e medimos as experiências de estigma e o funcionamento global. A frequência das respostas da secção de estigma foi semelhante à dos restantes estudos utilizando a CESQ. A frequência das respostas na secção de discriminação foi ligeiramente inferior à reportada noutros estudos. Verificámos a existência de uma associação entre a pontuação da subescala de discriminação, o sexo masculino e o facto de se viver na comunidade. A pontuação da subescala de discriminação está também correlacionada de forma positiva com o funcionamento global. Os alfas de Cronbach para a CESQ e para as suas subescalas foram considerados bons. Os coeficientes de correlação intraclasse foram igualmente considerados igualmente bons. Utilizando técnicas de análise fatorial, verificámos que a maior parte dos itens da CESQ se enquadrava em dois fatores, correspondendo sensivelmente às subescalas definidas previamente. Concluímos que o presente estudo explorou com sucesso a questão do estigma em Portugal, contribuindo em simultâneo para a validação do “Consumer Experiences of Stigma Questionnaire.--------------ABSTRACT: We reviewed the evolution of concept of stigma, its correlates and consequences, and analysed psychometric instruments that were used to study personal experiences of stigma. We provided an insight over research of stigma in Portugal. We reviewed relevant studies that use Consumer Experiences of Stigma Questionnaire and documented psychometric properties of this instrument. Our study aimed both to explore experiences of stigma in a Portuguese sample of people with severe mental illness and to contribute to the assessment of the psychometric properties of Consumer Experiences of Stigma Questionnaire and to the validation of its Portuguese translation. We performed a cross sectional descriptive and analytic study, collected socio-demographic data and measured experiences of stigma and global functioning. Frequency of responses regarding stigma section of CESQ matched previous studies using that scale. Frequency of responses in discrimination section was slightly lower than previously reported studies. We found an association between the discrimination score of CESQ and both male gender and living in the community. The discrimination score also positively correlated with global functioning. Cronbach alphas for CESQ and its subscales were good. Intraclass correlation coefficients for CESQ and stigma subscale were also good. Using factor analysis we found most of the items in CESQ would fit 2 factors, grossly corresponding to the previously defined subscales. We conclude that this study successfully explored stigma in Portugal, contributing in simultaneous to the validation of Consumer Experiences Questionnaire.

Predictive analytical and numerical modeling for orthogonal cutting

In this thesis, a predictive analytical and numerical modeling approach for the orthogonal cutting process is proposed to calculate temperature distributions and subsequently, forces and stress distributions. The models proposed include a constitutive model for the material being cut based on the work of Weber, a model for the shear plane based on Merchants model, a model describing the contribution of friction based on Zorev’s approach, a model for the effect of wear on the tool based on the work of Waldorf, and a thermal model based on the works of Komanduri and Hou, with a fraction heat partition for a non-uniform distribution of the heat in the interfaces, but extended to encompass a set of contributions to the global temperature rise of chip, tool and work piece. The models proposed in this work, try to avoid from experimental based values or expressions, and simplifying assumptions or suppositions, as much as possible. On a thermo-physical point of view, the results were affected not only by the mechanical or cutting parameters chosen, but also by their coupling effects, instead of the simplifying way of modeling which is to contemplate only the direct effect of the variation of a parameter. The implementation of these models was performed using the MATLAB environment. Since it was possible to find in the literature all the parameters for AISI 1045 and AISI O2, these materials were used to run the simulations in order to avoid arbitrary assumption.

The use of analytical hierarchy process in spatial decision support system for land use management

Geographic information systems give us the possibility to analyze, produce, and edit geographic information. Furthermore, these systems fall short on the analysis and support of complex spatial problems. Therefore, when a spatial problem, like land use management, requires a multi-criteria perspective, multi-criteria decision analysis is placed into spatial decision support systems. The analytic hierarchy process is one of many multi-criteria decision analysis methods that can be used to support these complex problems. Using its capabilities we try to develop a spatial decision support system, to help land use management. Land use management can undertake a broad spectrum of spatial decision problems. The developed decision support system had to accept as input, various formats and types of data, raster or vector format, and the vector could be polygon line or point type. The support system was designed to perform its analysis for the Zambezi river Valley in Mozambique, the study area. The possible solutions for the emerging problems had to cover the entire region. This required the system to process large sets of data, and constantly adjust to new problems’ needs. The developed decision support system, is able to process thousands of alternatives using the analytical hierarchy process, and produce an output suitability map for the problems faced.

Tradução e validação para Portugal do Patient Activation Measure 13 : em pessoas com Diabetes Mellitus tipo 2

RESUMO - Introdução: A Diabetes Mellitus Tipo 2 (DM2) tem uma elevada prevalência em todo o mundo, com impacto significativo a nível de Saúde Pública, na vida dos doentes e nos custos que lhe são associados. O Patiente Activation Measure 13 (PAM13) é um questionário que possibilita a avaliação das crenças, conhecimentos, motivação e capacidades de uma pessoa em relação à sua saúde; pelo que a sua utilização na DM2 é pertinente. Objetivos: Traduzir o PAM 13, versão curta, para Português de Portugal; Estabelecer as propriedades psicométricas da versão Portuguesa do PAM 13 (PAM13-P); Validar a PAM 13-P num grupo de pessoas com DM2. Material e métodos: O processo de tradução e adaptação cultural do questionário foi composto pelas fases: 1-Tradução, 2-Reconciliação e síntese, 3-Back translation, 4-Rever e sintetizar a Back translation, 5-Harmonização, 6-Revisão do comité de peritos, 7-Cognitive debriefing e 8-Avaliação final. Para validar o PAM13-P realizou-se um estudo observacional transversal analítico com uma amostra de conveniência, de pessoas com DM2, seguidas na Associação Protectora dos Diabéticos de Portugal. O questionário foi de autopreenchimento e foi consultado o processo clínico para obtenção da HbA1c. O tratamento estatístico foi realizado através do SPSS 21® e Winsteps v3.8.1®. Resultados e discussão: O processo de tradução e adaptação cultural foi realizado de acordo com as guidelines. Foram realizados 3 painéis de e-Delphi, com 21 participantes de áreas distintas, tendo-se obtido bons níveis de concordância. As principais modificações realizadas ao questionário foram a simplificação da linguagem e dos itens, obtendo-se as equivalências necessárias. O PAM13-P foi aplicado a 201 pessoas, sendo que a taxa de resposta foi de 83%. Na amostra analisada 57.3% eram homens. Obtiveram-se as médias de idade 67.1 anos e de duração de diabetes 17.3 anos. A média do score do PAM foi 58.5±10.1(41.8-90.5) pontos e 49,7% da amostra estava no nível 3 de ativação. Relativamente aos itens verificou-se que os itens mais difíceis foram o 13(56.1) e o 8(55.4) e o com menor dificuldade foi 4(38.5). As categorias de resposta tiveram um bom ajuste ao Modelo de Rasch. O ajustamento dos itens foi infit entre 0.779-1.177 e outfit entre 0.794-1.315. A fiabilidade dos indivíduos variou entre 0.77(real) e 0.83(modelo) e dos itens foi de 0.97 (real e modelo). O Alfa de Cronbach foi bom (α=0.82). Estas estatísticas foram semelhantes aos da validação do PAM13. Existiu uma relação entre o score do PAM e os itens de validação em 51%. Das variáveis analisadas, existia um relação do nível de ativação com a idade e com a HbA1c. Conclusões: O PAM13-P foi traduzido e adaptado culturalmente para Português e foi validado em pessoas com DM2, sendo as propriedades psicométricas boas.

Fx swaps and outright forward validation

UNL - NSBE

Development and validation of gold nanoprobes for human SNP detection towards commercial application

Conventional molecular techniques for detection and characterization of relevant nucleic acid (i.e. DNA) sequences are, nowadays, cumbersome, expensive and with reduced portability. The main objective of this dissertation consisted in the optimization and validation of a fast and low-cost colorimetric nanodiagnostic methodology for the detection of single nucleotide polymorphisms (SNPs). This was done considering SNPs associated to obesity of commercial interest for STAB VIDA, and subsequent evaluation of other clinically relevant targets. Also, integration of this methodology into a microfluidic platform envisaging portability and application on points-of-care (POC) was achieved. To warrant success in pursuing these objectives, the experimental work was divided in four sections: i) genetic association of SNPs to obesity in the Portuguese population; ii) optimization and validation of the non-cross-linking approach for complete genotype characterization of these SNPs; iii) incorporation into a microfluidic platform; and iv) translation to other relevant commercial targets. FTO dbSNP rs#:9939609 carriers had higher body mass index (BMI), total body fat mass, waist perimeter and 2.5 times higher risk to obesity. AuNPs functionalized with thiolated oligonucleotides (Au-nanoprobes) were used via the non-cross-linking to validate a diagnostics approach against the gold standard technique - Sanger Sequencing - with high levels of sensitivity (87.50%) and specificity (91.67%). A proof-of-concept POC microfluidic device was assembled towards incorporation of the molecular detection strategy. In conclusion a successful framework was developed and validated for the detection of SNPs with commercial interest for STAB VIDA, towards future translation into a POC device.