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.

Les Plus Anciens Mastodontes Tetralophodontes (Langhien inférieur Vb), évolution et remarques sur la Tetralophodontie

This study deals with mastodont teeth found near Lisbon in Lower Langhian (lower Middle Miocene) fluviatile, feldspathic sands (Vb division). Conclusions are as follows: 1. Tetralophodont molars (even if at a still primitive stade of the tetralophodont condition) do exist at least since lower Langhian times, and not only since late Middle Miocene as was previously known. 2. Tri- and tetralophodont structures may (and indeed do) coexist in the same individual: such examples do not correspond to transitional forms, but instead to a mosaic of juxtaposed characters (however this does not mean there are no transitional forms in other instances). 3. So these structures coexisted in a population not yet genetically separated beyond fertile cross-breeding, i.e. beyond species' level. 4. Origin of the tetralophodont molar was due to some mutation (s). but without crossing species, limits and even more genus'ones. 5. At this times probably soon after the first appearance of tetralophodont mutants, animals with such characters were a small but significant minority among the population (17% if account is taken on D4's: only 2% after M2's). 6. There was not then any direct and clear correlation between number of lophs (transversal crests) and tooth size, even if the increase of such number goes along with length's increase. 7. Dimensions (length in special) in tetralophodont teeth tend to exceed those in «normal» trilophodont teeth, this being particularly clear in D4, even if there is no clear distinction: the situation is quite the same, maybe less marked, with the M2. 8. According to the preceding conclusions there are no reasons to segregate different taxa among such mastodont population on the grounds of the presence in D4, M1 and M2 of 3 or 4 crests (this character being regarded as diagnostic of the genus Tetralophodon). 9. On the contrary, if any natural (in biological sense) classification is disregarded and a morphological parataxonomy is adopted there should be considered both Gomphotherium angustidens and Tetralophodon sp.: however this is absolutely not our opinion.

Notes sur la géologie et la paleontology du Miocène de Lisbonne; (XX) les plus anciens mastodontes tetralophodontes (Langhien inférieur Vb), évolution et remarques sur la tetralophodontie

This study deals with mastodont teeth found near Lisbon in Lower Langhian (lower Middle Miocene) fluviatile, feldspathic sands (Vb division). Conclusions are as follows: 1. Tetralophodont molars (even if at a still primitive stade of the tetralophodont condition) do exist at least since lower Langhian times, and not only since late Middle Miocene as was previously known. 2. Tri- and tetralophodont structures may (and indeed do) coexist in the same individual: such examples do not correspond to transitional forms, but instead to a mosaic of juxtaposed characters (however this does not mean there are no transitional forms in other instances). 3. So these structures coexisted in a population not yet geneticaliy separated beyond fertile cross-breeding, i.e. beyond species'level. 4. Origin of the tetralophodont molar was due to some mutation (s). but without crossing species, limits and even more genus' ones. 5. At this times probably soon after the first appearance of tetralophodont mutants, animals with such characters were a small but signifiant minority among the population (17% if account is taken on D4's: only 2% after M2's). 6. There was not then any direct and clear correlation between number of lophs (transversal crests) and tooth size, even if the increase of such number goes along with length's increase. 7. Dimensions (length in special) in tetralophodont teeth tend to exceed those in «normal» trilophodont teeth, this being particularly clear in D4, even if there is no clear distinction: the situation is quite the same, maybe less marked, with the M2. 8. According to the preceding conclusions there are no reasons to segregate different taxa among such mastodont population on the grounds of the presence in D4, M1 and M2 of 3 or 4 crests (this character being regarded as diagnostic of the genus Tetralophodon). 9. On the contrary, if any natural (in biological sense) classification is disregarded and a morphological parataxonomy is adopted there should be considered both Gomphotherium angustidens and Tetralophodon sp.: however this is absolutely not our opinion.

Origin and evolution of Upper Triassic to Miocene clay-mineral associations from the eastern Algarve of Portugal

XRD-analyses of pelitic deposits of Upper Jurassic to Miocene age occuring in the eastern Algarve (Portugal), give evidence of the occurrence of detrital clay minerals of continental origin as well as of conspicuous neoformations of marine provenance. The vertical succession of clay-mineral associations indicates the existence of three distinctive evolutionary cycles which are thought to reflect tectonically controlled transgressive-regressive events.

Wetland Habitat Studies using various Classification Techniques on Multi-Spectral Landsat Imagery: Case study: Tram chim National Park, Dong Thap Vietnam

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

Automatic sleep apnea detection and sleep classification using the ECG and the SpO2 signals

Dissertation for a Masters Degree in Computer and Electronic Engineering

An overview of the higher level classification of Pucciniomycotina based on combined analyses of nuclear large and small subunit rDNA sequences

Mycologia, Vol. 98, nº6

The contribution of multitemporal information from multispectral satellite images for automatic land cover classification at the national scale

Thesis submitted to the Instituto Superior de Estatística e Gestão de Informação da Universidade Nova de Lisboa in partial fulfillment of the requirements for the Degree of Doctor of Philosophy in Information Management – Geographic Information Systems

The detrital origin of the Moncorvo ordovician ironstones

The Moncorvo Ordovician ironstones in northeastern Portugal consist of iron ore sedimentary horizons frequently interbanded with psamites and quartzites. Ore reserves may probably exceed 1 000 million tonnes and this makes Moncorvo the largest iron ore deposit in the European Union. Compact poorly banded massive layers may exceed 90 meters in thickness which is quite an extraordinary feature for a Phanerozoic deposit. If the thickness of Precambrian deposits may reach a few hundred meters, the thickness of Phanerozoic deposits never exceed a maximum of 15 meters generally forming a number of comparatively thin layers confined to a particular member of a sedimentary sequence. A detailed microscopic analysis of the ores revealed that initially a compact magnetite/quartzite layer, detrital in character (the magnetite occasionally showing chromite cores), was deposited by entrapment in near shore lagoons where rivers debouched, rather than in the open sea. This stage was followed by oscilating and transgressive shore lines which gave rise to breaks in sedimentation in combined river delta and shallow water marine environment where detrital material and fine iron oxide and clay suspensions were deposited in fluctuating environments. These events gave rise to layers of both magnetite (martite) and specularite intergrown with quartz, silicates and phosphates. Textural and mineralogical studies show that the deposits consist of ferruginous clastic sediments and are not chemically deposited cherts. Field, geological and palaeontological evidence also supports a detrital origin, the facies being typical of zones rich in oxygen and close to the feeding continent. The uncommon huge development of Moncorvo was due to the fact that the deposits occur in restricted basins on a continental platform were clastic sediments were predominantly deposited. Not only morphologically but also chemically the deposits are more similar to Precambrian iron formations than to Phanerozoic ironstones.

Unsupervised automatic music genre classification

Trabalho apresentado no âmbito do Mestrado em Engenharia Informática, como requisito parcial para obtenção do grau de Mestre em Engenharia Informática

Automatic cymbal classification

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

The origin and spread of the war chariot

Dissertação apresentada para cumprimento dos requisitos necessários à obtenção do grau de Mestre em História - Civilizações do Médio Oriente e Ásia Antiga

Multi-scale vegetation classification using earth observation data of the Sundarban mangrove forest, Bangladesh.

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

Paralisia cerebral na Região de Lisboa e Vale do Tejo : factores de risco : caracterização multidimensional : inclusão escolar

RESUMO - A Paralisia Cerebral (PC) deve ser olhada como uma patologia do neurodesenvolvimento: a infância é um período de actividade exploratória por essência, a restrição motora condiciona as várias áreas do desenvolvimento. Contextos, apoios, oportunidades e experiências de vida serão determinantes no desenvolvimento de todo o seu potencial. Objectivos/finalidade: Identificar, descrever, comparar e analisar factores de risco associados à PC, sua caracterização multidimensional e integração escolar aos 5 e 10 anos. Procurou‐se contribuir para a sua prevenção primária e secundária, e obter dados para planeamento e implementação dos programas de apoio. Métodos: Adoptou‐se a abordagem do Programa Nacional de Vigilância da Paralisia Cerebral (PNVPC) e da Surveillance of Cerebral Palsy in Europe (SCPE). Analisaram‐se factores de risco, competências funcionais, défices associados, severidade e integração escolar de duas populações de Lisboa e Vale do Tejo, (nascimento 1996/1997‐2001/2002 e prevalência aos 5 e 10 anos). Descreveram‐se os dados, efectuaram‐se correlações, aplicaram‐se testes de independência e compararam‐se com dados dos nadovivos, dados nacionais e europeus. Analisaram‐se os factores que influenciaram a integração escolar através de métodos de regressão logística. Resultados/Conclusões/Recomendações: 1,65‰ e 1,57‰ dos nadovivos desenvolveram PC; a prevalência aos 5 anos foi de 1,7‰ e de 1,48‰; 5,9% e 7,9% faleceram antes dos 5 anos. Em 2001/2002 verificou‐se aumento de: PC espástica bilateral‐2/3membros, prematuridade, causa pos‐neonatal, níveis funcionais ligeiros e graves; percentil estaturo‐ponderal <3 (5‐anos). Diminuição de: disquinésia, anóxia e alguns défices associados. Destacaram‐se as associação: prematuridade e PC espástica bilateral‐ 2/3membros; nascer de termo e anóxia, disquinésia, primíparas, défices associados e severidade; infecção pré‐natal e QI<50, epilepsia e severidade; causa pos‐neonatal e PC espástica bilateral‐4membros e múltiplos défices. Aos 5 anos, as variáveis explicativas para a não inclusão escolar foram: QI<50 e epilepsia; uma elevada percentagem de crianças com PC moderada/grave encontrava‐se integrada; 75% das que se encontravam desintegradas mantiveram‐se nesta situação aos 10. Nesta idade, as variáveis explicativas para a não inclusão escolar foram: QI<50 e motricidade fina; 35,1% encontrava‐se fora do ensino regular; 4,5%, embora em idade de escolaridade obrigatória, não frequentavam qualquer estabelecimento escolar. Informação sistematizada, abrangente, objectiva, simples e acessível, sobre novos casos de PC, factores de risco, prevalência em idades‐chave e caracterização multidimensional constitui uma ferramenta clínica e epidemiológica, que deve sustentar as políticas de saúde, educacionais e sociais, contribuindo para a permanência destas crianças no ensino regular, trazendo às crianças e famílias o suporte que as encorajem e sustentem nestes processos. ABSTRACT ------- Cerebral Palsy (CP) must be recognized as a neurodevelopmental disorder: childhood is, on its nature, a period for exploring the environment and therefore motor deficit interferes with all developmental areas. The context, support, opportunities and life experiences are determinants for the development of his full potential. Objective/Aim: To identify, describe, compare and analyze CP risk factors the multidimensional characterization and school integration levels at the age of 5 and 10 years. We aim to contribute to CP primary and secondary prevention and provide information for service planning and implementation of support programs. Methods: The approach of National Cerebral Palsy Surveillance Programme (NCPSP) and Surveillance of Cerebral Palsy in Europe (SCPE) were used. For two groups of children from Lisboa e Vale do Tejo region, birth data 1996/1997‐2001/2002 and prevalence at 5 and 10 years, were analyzed: CP risk factors, functional ability, associated impairments, severity and school integration settings. Data collected was described, analyzed using correlations, applied tests of independence and compared with new born data, national data and european data. To analyze the factors related to school inclusive settings, logistic regression was appealed. Results/Conclusions/Recommendations: 1,65% and 1,57‰ of the new‐born alive developed CP. The prevalence at 5 years was 1,7‰ and 1,48‰ 5,9% and 7,9% died before their 5th birthday. Bilateral spastic CP 2/3limb, preterm birth, cases of post‐neonatal origin, mild and sever functional impairment; weight and height percentile <3 at 5 years old Increased in 2001. Decreased dyskinetic CP, anoxia and some additional imparments. Were identified among other the association between prematurity and spastic bilateral CP‐2/3 members; born at term and anoxia, dyskinetic CP, first child, associated impairments and severity; prenatal infection and IQ<50, epilepsy and severity; post‐neonatal cause and spastic bilateral CP‐4 members and associated impairments. At 5‐years‐old the more explanatory variables for not be in a school inclusive settings were IQ<50 and epilepsy, a high percentage of children with moderate/severe CP was attending regular school, but most children who were out of inclusive settings at 5 years continue on this situations at 10‐years‐old. At this age the more explanatory variables for not be in a school inclusive settings were: IQ<50 and upper limb function; 35,1% were out of regular school; 4,5%, even in compulsory school age, are out of school. Standardized comprehensive, objective, simple and accessible information about CP new cases, risk factors, prevalence in the key‐age and children multidimensional characterization constitutes a clinical and epidemiological tool that should sustain health, educational and social policy. This would support the continuity of these children in regular schools, encouraging g them and their families in these processes.

Unsupervised classification of remote sensing images combining Self Organizing Maps and segmentation techniques

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