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.

Molecular diversity of bla genes in Klebsiella pneumoniae and Escherichia coli isolates

Dissertation presented to obtain a Ph.D. degree in Biology, speciality Microbiology, by Universidade Nova de Lisboa, Faculdade de Ciências e Tecnologia

Influence of constraining and confinement in the molecular mobility of low molecular weight materials

Dissertation presented to obtain a Ph.D. Degree in Chemical Physics

Dielectric study of triton X100: a glass-forming liquid

Dissertação para obtenção do grau de mestre em Engenharia de Materiais

Application of NMR RDC’s, relaxation and self-diffusion for the study of dynamic processes of small molecules in solution

Spin-lattice Relaxation, self-Diffusion coefficients and Residual Dipolar Couplings (RDC’s) are the basis of well established Nuclear Magnetic Resonance techniques for the physicochemical study of small molecules (typically organic compounds and natural products with MW < 1000 Da), as they proved to be a powerful and complementary source of information about structural dynamic processes in solution. The work developed in this thesis consists in the application of the earlier-mentioned NMR techniques to explore, analyze and systematize patterns of the molecular dynamic behavior of selected small molecules in particular experimental conditions. Two systems were chosen to investigate molecular dynamic behavior by these techniques: the dynamics of ion-pair formation and ion interaction in ionic liquids (IL) and the dynamics of molecular reorientation when molecules are placed in oriented phases (alignment media). The application of NMR spin-lattice relaxation and self-diffusion measurements was applied to study the rotational and translational molecular dynamics of the IL: 1-butyl-3-methylimidazolium tetrafluoroborate [BMIM][BF4]. The study of the cation-anion dynamics in neat and IL-water mixtures was systematically investigated by a combination of multinuclear NMR relaxation techniques with diffusion data (using by H1, C13 and F19 NMR spectroscopy). Spin-lattice relaxation time (T1), self-diffusion coefficients and nuclear Overhauser effect experiments were combined to determine the conditions that favor the formation of long lived [BMIM][BF4] ion-pairs in water. For this purpose and using the self-diffusion coefficients of cation and anion as a probe, different IL-water compositions were screened (from neat IL to infinite dilution) to find the conditions where both cation and anion present equal diffusion coefficients (8% water fraction at 25 ºC). This condition as well as the neat IL and the infinite dilution were then further studied by 13C NMR relaxation in order to determine correlation times (c) for the molecular reorientational motion using a mathematical iterative procedure and experimental data obtained in a temperature range between 273 and 353 K. The behavior of self-diffusion and relaxation data obtained in our experiments point at the combining parameters of molar fraction 8 % and temperature 298 K as the most favorable condition for the formation of long lived ion-pairs. When molecules are subjected to soft anisotropic motion by being placed in some special media, Residual Dipolar Couplings (RDCs), can be measured, because of the partial alignment induced by this media. RDCs are emerging as a powerful routine tool employed in conformational analysis, as it complements and even outperforms the approaches based on the classical NMR NOE or J3 couplings. In this work, three different alignment media have been characterized and evaluated in terms of integrity using 2H and 1H 1D-NMR spectroscopy, namely the stretched and compressed gel PMMA, and the lyotropic liquid crystals CpCl/n-hexanol/brine and cromolyn/water. The influence that different media and degrees of alignment have on the dynamic properties of several molecules was explored. Different sized sugars were used and their self-diffusion was determined as well as conformation features using RDCs. The results obtained indicate that no influence is felt by the small molecules diffusion and conformational features studied within the alignment degree range studied, which was the 3, 5 and 6 % CpCl/n-hexanol/brine for diffusion, and 5 and 7.5 % CpCl/n-hexanol/brine for conformation. It was also possible to determine that the small molecules diffusion verified in the alignment media presented close values to the ones observed in water, reinforcing the idea of no conditioning of molecular properties in such media.

Development and modulation of magnetic responsive supported ionic liquid membranes

The work presented in this thesis aims at developing a new separation process based on the application of supported magnetic ionic liquid membranes, SMILMs, using magnetic ionic liquids, MILs. MILs have attracted growing interest due to their ability to change their physicochemical characteristics when exposed to variable magnetic field conditions. The magnetic responsive behavior of MILs is thus expected to contribute for the development of more efficient separation processes, such as supported liquid membranes, where MILs may be used as a selective carrier. Driven by the MILs behavior, these membranes are expected to switch reversibly their permeability and selectivity by in situ and non-invasive adjustment of the conditions (e.g. intensity, direction vector and uniformity) of an external applied magnetic field. The development of these magnetic responsive membrane processes were anticipated by studies, performed along the first stage of this PhD work, aiming at getting a deep knowledge on the influence of magnetic field on MILs properties. The influence of the magnetic field on the molecular dynamics and structural rearrangement of MILs ionic network was assessed through a 1H-NMR technique. Through the 1H-NMR relaxometry analysis it was possible to estimate the self-diffusion profiles of two different model MILs, [Aliquat][FeCl4] and [P66614][FeCl4]. A comparative analysis was established between the behavior of magnetic and non-magnetic ionic liquids, MILs and ILs, to facilitate the perception of the magnetic field impact on MILs properties. In contrast to ILs, MILs show a specific relaxation mechanism, characterized by the magnetic dependence of their self-diffusion coefficients. MILs self-diffusion coefficients increased in the presence of magnetic field whereas ILs self-diffusion was not affected. In order to understand the reasons underlying the magnetic dependence of MILs self-diffusion, studies were performed to investigate the influence of the magnetic field on MILs’ viscosity. It was observed that the MIL´s viscosity decreases with the increase of the magnetic field, explaining the increase of MILs self-diffusion according to the modified Stokes- Einstein equation. Different gas and liquid transport studies were therefore performed aiming to determine the influence of the magnetic behavior of MILs on solute transport through SMILMs. Gas permeation studies were performed using pure CO2 andN2 gas streams and air, using a series of phosphonium cation based MILs, containing different paramagnetic anions. Transport studies were conducted in the presence and absence of magnetic field at a maximum intensity of 1.5T. The results revealed that gas permeability increased in the presence of the magnetic field, however, without affecting the membrane selectivity. The increase of gas permeability through SMILMs was related to the decrease of the MILs viscosity under magnetic field conditions.(...)

A minimal version of a Protein Tyrosine Phosphatase

Phosphatase and tensin homologue (PTEN) protein belongs to the family of protein tyrosine phos-phatase. Mutations on the phosphatase and tensin homologue (PTEN) protein are highly observed in diverse types of human tumors, being mostly identified on the phosphatase domain of the protein. Although PTEN is a modular protein composed by a phosphatase domain and a C2 domain for mem-brane anchoring, this work aimed at developing a minimal version of PTEN´s phosphatase domain. The minimal version (Small Domain) comprises a 28 residue peptide, with the PTEN 8-mer catalytic peptide accommodated between a α-helix and β-turn as observed in PTEN native structure. Firstly, a de novo prediction of the Small Domain´s secondary structure was carried out by molecular modeling tools. The stability of the predicted structures were then evaluated by Molecular Dynamics. Automated molecular docking of PTEN natural substrate PIP3, its analogue (Inositol) and a PTEN inhibitor (L-tar-tare) were performed with the modeled structure, and PTEN used as a positive control. The gene en-coding for Small Domain was designed and cloned into an expression vector at N-terminal of Green Fluorescence Protein (GFP) encoding gene. The fusion protein was then expressed in Escherichia coli cells. Different expression conditions have been explored for the production of the fusion protein to minimize the formation of inclusion bodies.