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.

Run-time monitoring approach for the shark kernel

Typically common embedded systems are designed with high resource constraints. Static designs are often chosen to address very specific use cases. On contrast, a dynamic design must be used if the system must supply a real-time service where the input may contain factors of indeterminism. Thus, adding new functionality on these systems is often accomplished by higher development time, tests and costs, since new functionality push the system complexity and dynamics to a higher level. Usually, these systems have to adapt themselves to evolving requirements and changing service requests. In this perspective, run-time monitoring of the system behaviour becomes an important requirement, allowing to dynamically capturing the actual scheduling progress and resource utilization. For this to succeed, operating systems need to expose their internal behaviour and state, making it available to the external applications, usually using a run-time monitoring mechanism. However, such mechanism can impose a burden in the system itself if not wisely used. In this paper we explore this problem and propose a framework, which is intended to provide this run-time mechanism whilst achieving code separation, run-time efficiency and flexibility for the final developer.

Phex: a computational tool for plate heat exchanges design problems

This paper presents a computational tool (PHEx) developed in Excel VBA for solving sizing and rating design problems involving Chevron type plate heat exchangers (PHE) with 1-pass-1-pass configuration. The rating methodology procedure used in the program is outlined, and a case study is presented with the purpose to show how the program can be used to develop sensitivity analysis to several dimensional parameters of PHE and to observe their effect on transferred heat and pressure drop.

Advances and challenges in oral health after a decade of the “Smiling Brazil” Program

ABSTRACT OBJECTIVE To analyze oral health work changes in primary health care after Brazil’s National Oral Health Policy Guidelines were released. METHODS A literature review was conducted on Medline, LILACS, Embase, SciELO, Biblioteca Virtual em Saúde, and The Cochrane Library databases, from 2000 to 2013, on elements to analyze work changes. The descriptors used included: primary health care, family health care, work, health care policy, oral health care services, dentistry, oral health, and Brazil. Thirty-two studies were selected and analyzed, with a predominance of qualitative studies from the Northeast region with workers, especially dentists, focusing on completeness and quality of care. RESULTS Observed advances focused on educational and permanent education actions; on welcoming, bonding, and accountability. The main challenges were related to completeness; extension and improvement of care; integrated teamwork; working conditions; planning, monitoring, and evaluation of actions; stimulating people’s participation and social control; and intersectorial actions. CONCLUSIONS Despite the new regulatory environment, there are very few changes in oral health work. Professionals tend to reproduce the dominant biomedical model. Continuing efforts will be required in work management, training, and permanent education fields. Among the possibilities are the increased engagement of managers and professionals in a process to understand work dynamics and training in the perspective of building significant changes for local realities.

Towards a GIS-T database design and implementation for public transit planning: the case study of Dar-Es-salaam metropolitan city, Tanzania.

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

Benchmarking de identificação de ano: criação de um site para benchmarking de técnicas de identificação de dano

Dissertação para obtenção do grau de Mestre em Engenharia Mecânica na Área de Manutenção e Produção

CodeSkelGen: a program skeleton generator

Existent computer programming training environments help users to learn programming by solving problems from scratch. Nevertheless, initiating the resolution of a program can be frustrating and demotivating if the student does not know where and how to start. Skeleton programming facilitates a top-down design approach, where a partially functional system with complete high level structures is available, so the student needs only to progressively complete or update the code to meet the requirements of the problem. This paper presents CodeSkelGen - a program skeleton generator. CodeSkelGen generates skeleton or buggy Java programs from a complete annotated program solution provided by the teacher. The annotations are formally described within an annotation type and processed by an annotation processor. This processor is responsible for a set of actions ranging from the creation of dummy methods to the exchange of operator types included in the source code. The generator tool will be included in a learning environment that aims to assist teachers in the creation of programming exercises and to help students in their resolution.

Design and implementation of an autonomous, proactive, and reactive software infrastructure to help improving the management level of projects

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

Conceção e implementação de um programa de melhoria da qualidade: estudo de caso na Eurico Ferreira, S.A.

Os programas de melhoria contínua dos processos são cada vez mais a aposta das empresas para fazer face ao mercado. Através da implementação destes programas é possível conferir simplicidade e padronização aos processos e consequentemente reduzir os custos com desperdícios internos relacionados com a qualidade dos mesmos. As ferramentas de melhoria da qualidade e as ferramentas associadas ao Lean Thinking representam um pilar importante no sucesso de qualquer programa de melhoria contínua dos processos. Estas ferramentas constituem meios úteis na análise, controlo, organização de dados importantes para a correta tomada de decisão nas organizações. O presente projeto tem como principal objetivo a conceção e implementação de um programa de melhoria da qualidade na Eurico Ferreira, S.A., tendo por base a avaliação da satisfação do cliente e a aplicação dos 5S. Neste contexto, o trabalho teve como fundamentação teórica a Gestão da Qualidade, Lean Thinking e algumas ferramentas de ambas as matérias. Posteriormente foi selecionada a área de negócio da empresa a abordar. Após a seleção, realizou-se um diagnóstico inicial do processo identificando os diversos pontos de melhoria onde foram aplicadas algumas ferramentas do Lean Thinking, nomeadamente o Value Stream Mapping e a metodologia 5S. Com a primeira foi possível construir um mapa do estado atual do processo, no qual estavam representados todos os intervenientes assim como o fluxo de materiais e de informação ao longo do processo. A metodologia 5S permitiu atuar sobre os desperdícios, identificando e implementando diversas melhorias no processo. Concluiu-se que a implementação das ferramentas contribuiu eficientemente para a melhoria contínua da qualidade nos processos, tendo sido decisão da coordenação alargar o âmbito do projeto aos restantes armazéns do centro logístico da empresa. Pode afirmar-se com recurso à satisfação do cliente expressa através da evolução favorável do Service-level agreement que as ferramentas implementadas têm gerado resultados muito positivos no curto prazo.

Planning smart cities using wireless low energy monitoring systems

Thesis to obtain the Master Degree in Electronics and Telecommunications Engineering

Modelação do comportamento dinâmico de sistemas barragem-fundação-albufeira: análise sísmica de uma barragem de abóbadas múltiplas

Trabalho Final de Mestrado elaborado no Laboratório Nacional de Engenharia Civil (LNEC) para a obtenção do grau de Mestre em Engenharia Civil pelo Instituto Superior de Engenharia de Lisboa no âmbito do protocolo de cooperação entre o ISEL e o LNEC

Lean and quality management integration: improvement program implementation in a service company

Lean Thinking is an important pillar in the success of any program of continuous improvement process. Its tools are useful means in the analysis, control and organization of important data for correct decision making in organizations. This project had as main objective the design of a program of quality improvement in Eurico Ferreira, S.A., based on the evaluation of customer satisfaction and the implementation of 5S. Subsequently, we have selected which business area of the company to address. After the selection, there was an initial diagnostic procedure, identifying the various points of improvement to which some tools of Lean Thinking have been applied, in particular Value Stream Mapping and 5S methodology. With the first, we were able to map the current state of the process in which all stakeholders were represented as well as the flow of materials and information throughout the process. The 5S methodology allowed to act on the wastage, identifying and implementing various process improvements.

Optimizing potentiometric ionophore and electrode design for environmental on-site control of antibiotic drugs: Application to sulfamethoxazole

Potentiometric sensors are typically unable to carry out on-site monitoring of environmental drug contaminants because of their high limits of detection (LODs). Designing a novel ligand material for the target analyte and managing the composition of the internal reference solution have been the strategies employed here to produce for the first time a potentiometric-based direct reading method for an environmental drug contaminant. This concept has been applied to sulfamethoxazole (SMX), one of the many antibiotics used in aquaculture practices that may occur in environmental waters. The novel ligand has been produced by imprinting SMX on the surface of graphitic carbon nanostructures (CN) < 500 nm. The imprinted carbon nanostructures (ICN) were dispersed in plasticizer and entrapped in a PVC matrix that included (or not) a small amount of a lipophilic additive. The membrane composition was optimized on solid-contact electrodes, allowing near-Nernstian responses down to 5.2 μg/mL and detecting 1.6 μg/mL. The membranes offered good selectivity against most of the ionic compounds in environmental water. The best membrane cocktail was applied on the smaller end of a 1000 μL micropipette tip made of polypropylene. The tip was then filled with inner reference solution containing SMX and chlorate (as interfering compound). The corresponding concentrations were studied for 1 × 10−5 to 1 × 10−10 and 1 × 10−3 to 1 × 10−8 mol/L. The best condition allowed the detection of 5.92 ng/L (or 2.3 × 10−8 mol/L) SMX for a sub-Nernstian slope of −40.3 mV/decade from 5.0 × 10−8 to 2.4 × 10−5 mol/L.

A Novel Run-Time Monitoring Architecture for Safe and Efficient Inline Monitoring

20th International Conference on Reliable Software Technologies - Ada-Europe 2015 (Ada-Europe 2015), Madrid, Spain.

A disposable glass-based immunosensor for monitoring the cancer biomarker CEA in urine

NanoPT 2014 International Conference, in Portugal, on February 12-14. Poster presentation based on topic Nanobio/Nanomedicine