The purpose of this research was to conduct an exploratory study into the similarity of motivations of conformers and anti-conformers, particularly in relation to the personality traits, need for affiliation, dependency needs, and anxiety.

In an effort to determine if there is a specific conformational structure which is most effective at the appropriate active physiological site, the synthesis of a group of sterically restricted analogs was undertaken. A portion of the polymethylene carbon skeleton of glutaric acid was replaced by selected aromatic carbons in benzenedicarboxylic acids to produce a series of ridged conformers, and the relative plant growth regulating properties of these derivatives were determined.

In this work we address two problems in computational chemistry relevant to biomolecular modeling. In the first project, we consider the conformer space of melatonin as a a representative example of “real-life” flexible biomolecules. Geometries for all 52 unique conformers are optimized using spin-component scaled MP2, and then relative energies are obtained at the CCSD (T) level near the complete basis set limit. These are then used to validate a variety of DFT methods with and without empirical dispersion corrections, as well as some lower-level ab initio methods. Basis set convergence is found to be relatively slow due to internal C-H…O and C-H…N contacts. Absent dispersion corrections, many DFT functionals will transpose the two lowest conformers. Dispersion corrections resolve the problem for most functionals. Double hybrids yield particularly good performance, as does MP2.5. In the second project, we propose a simple DFT-based diagnostic for nondynamical correlation effects. Aλ= (1-TAE [ΧλC]/TAE[XC])/λ where TAE is the total atomization energy, XC the “pure” DFT exchange-correlation functional, and ΧλC the corresponding hybrid with 100λ% HF-type exchange. The diagnostic is a good predictor for sensitivity of energetics to the level of theory, unlike most of the wavefunction-based diagnostics. For GGA functionals, Aλ values approaching unity indicate severe non-dynamical correlation. The diagnostic is only weakly sensitive to the basis set (beyond polarized double zeta) and can be applied to problems beyond practical reach of wavefunction ab-initio methods required for other diagnostics.

RNA backbone is biologically important with many roles in reactions and interactions, but has historically been a challenge in structural determination. It has many atoms and torsions to place, and often there is less data on it than one might wish. This problem leads to both random and systematic error, producing noise in an already high-dimensional and complex distribution to further complicate data-driven analysis. With the advent of the ribosomal subunit structures published in 2000, large RNA structures at good resolution, it became possible to apply the Richardson laboratory's quality-filtering, visualization, and analysis techniques to RNA and develop new tools for RNA as well. A first set of 42 RNA backbone rotamers was identified, developed, and published in 2003; it has since been thoroughly overhauled in conjunction with the backbone group of the RNA Ontology Consortium to combine the strengths of different approaches, incorporate new data, and produce a consensus set of 46 conformers. Meanwhile, extensive work has taken place on developing validation and remodeling tools to correct and improve existing structures as well as to assist in initial fitting. The use of base-phosphate perpendicular distances to identify sugar pucker has proven very useful in both hand-refitting and the semi-automated process of using RNABC (RNA Backbone Correction), a program developed in conjunction with Dr. Jack Snoeyink's laboratory. The guanine riboswitch structure ur0039/1U8D, by Dr. Rob Batey's laboratory, has been collaboratively refit and rerefined as a successful test case of the utility of these tools and techniques. Their testing and development will continue, and they are expected to help to improve RNA structure determination in both ease and quality. / Dissertation

This thesis discusses Kaplan, a free conformer searching package, available at github.com/PeaWagon/Kaplan / Conformer searching algorithms find minima in the Potential Energy Surface (PES) of a molecule, usually by following a torsion-driven approach. The minima represent conformers, which are interchangeable via free rotation around bonds. Conformers can be used as input to computational analyses, such as drug design, that can convey molecular reactivity, structure, and function. With an increasing number of rotatable bonds, finding optima in the PES becomes more complicated, as the dimensionality explodes. Kaplan is a new, free and open-source software package written by the author that uses a ring-based Evolutionary Algorithm (EA) to find conformers. The ring, which contains population members (or pmems), is designed to allow initial PES exploration, followed by exploitation of individual energy wells, such that the most energetically-favourable structures are returned. The strengths and weaknesses of existing publicly available conformer searchers are discussed, including Balloon, RDKit, Openbabel, Confab, Frog2, and Kaplan. Since RDKit is usually considered to be the best free package for conformer searching, its conformers for the amino acids were optimised using the MMFF94 forcefield and compared to the conformers generated by Kaplan. Amino acid conformers are well characterised, and provide insight for protein substructure. Of the 20 molecules, Kaplan found a lower energy minima for 12 of the structures and tied for 5 of them. Kaplan allows the user to specify which dihedrals (by atom indices) to optimise and angles to use, a feature that is not offered by other programs. The results from Kaplan were compared to a known dataset of amino acid conformers. Kaplan identified all 57 conformers of methionine to within 1.2Å, and found identical conformers for the 5 lowest-energy structures (i.e. within 0.083Å), following forcefield optimisation. / Thesis / Master of Science (MSc) / A conformer search affords the low-energy arrangements of atoms that can be obtained via rotation around bonds. Conformers provide insight about the chemical reactivity and physical properties of a molecule. With increasing molecule size, the number of possible conformers increases exponentially. To search the space of possible conformers, this thesis presents Kaplan, which is a software package that implements a novel directed, stochastic, sampling technique based on an Evolutionary Algorithm (EA). Kaplan uses a special type of EA that stores sets of conformers in a ring-based structure. Unlike other conformer-specific packages, Kaplan provides the means to analyse and interact with found conformers. Known conformers of amino acids are used to verify Kaplan. Other tools for generating conformers are discussed, including a comparison of freely available software. Kaplan effectively finds the conformers of small molecules, but requires additional parametrisation to find the conformers of mid-sized molecules, such as Penta-Alanine.

Submitted by JÚLIO HEBER SILVA (julioheber@yahoo.com.br) on 2017-06-26T18:28:31Z No. of bitstreams: 2 Dissertação - Rafael Carvalho Couto - 2013.pdf: 41193945 bytes, checksum: 74a020dad23640afb84a085b841b91aa (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Approved for entry into archive by Cláudia Bueno (claudiamoura18@gmail.com) on 2017-07-07T20:26:09Z (GMT) No. of bitstreams: 2 Dissertação - Rafael Carvalho Couto - 2013.pdf: 41193945 bytes, checksum: 74a020dad23640afb84a085b841b91aa (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2017-07-07T20:26:10Z (GMT). No. of bitstreams: 2 Dissertação - Rafael Carvalho Couto - 2013.pdf: 41193945 bytes, checksum: 74a020dad23640afb84a085b841b91aa (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2013-04-15 / This work focuses on the development of the software MGA, which aims to determine the lowest energy structures of a given molecular system, using Genetic Algorithm (GA). The GA is a method of artificial intelligence that was developed to work with finding the best solutions of the specified conditions, ie, an algorithm that seeks the best answer desired, an optimal result. The MGA uses three techniques: Random Search (RS), Noninclusive Genetic Algorithm (NGA), Inclusive Genetic Algorithm (IGA). The last one is characterized by a new type of evolutionary strategy that allows in a single calculation and a single cycle, obtain several minimum of the potential energy surface. For optimum operation of the algorithm, was made an optimization of the parameters used in MGA, through response surface methodology. Using the techniques RS, IGA and NGA, were determined 141 distinct molecular structures of the amino acid asparagine. In the electronic structure calculations were considered the semi-empirical methods PM3, AM1 and RM1; and DFT potentials, with basis sets 6-311G ** and PC1. The RS determined the Global Minimum (GM) with ease, for the different potentials used, and proved that it’s quite useful in determining molecular geometries where there is no accuracy in the determination of local minima in order of energy. The NGA is efficient in determining the GM, performing in a shorter time, if compared to RS and IGA. The IGA proved to be a more robust method than the others, because in addition to determining the GM, it can find the local minima in order of energy. Performing calculations on an intermediate time of RS and NGA, the IGA determined the GM as the NGA, and found structures that were not founded using RS. The GM’s of asparagine determined using the potentials PC1, PM3, AM1 and RM1 have a large structural difference. This demonstrates that different potencials used in the electronic structure calculations may lead to different results. By analyzing the structures obtained for potentials PC1, PM3, AM1 and RM1, using the IGA, it appears that there is a difference in the topology of the potential energy surface of these potentials. / O presente trabalho é focado no desenvolvimento do software MGA, que tem como objetivo a determinação das estruturas de menor energia de um dado sistema molecular, utilizando o Algoritmo Genético (AG). O AG é um método de inteligência artificial que foi desenvolvido para trabalhar com a procura de soluções que melhor atendam as condições especificadas, isto é, um algoritmo que procura a melhor resposta desejada, um resultado ótimo. O MGA utiliza três técnicas: Busca Aleatória (RS), Algoritmo Genético Não-inclusivo (NGA), Algoritmo Genético Inclusivo (IGA). Este último é caracterizado por um novo tipo de estratégia evolutiva que permite em um único cálculo e um único ciclo evolucionário obter diversos mínimos da superfície de energia potencial. Para o melhor funcionamento do algoritmo, foi feita uma otimização dos parâmetros utilizados do MGA, através da metodologia de superfície de resposta. Utilizando as técnicas RS, NGA e IGA, foram determinadas 141 estruturas moleculares distintas do aminoácido asparagina. Nos cálculos de estrutura eletrônica foram considerados os métodos semi-empíricos PM3, AM1 e RM1; e potenciais DFT, com os conjuntos de base 6-311G** e PC1. O RS determinou o Mínimo Global (GM) com facilidade, para os diferentes potenciais utilizados, e se mostrou bastante útil na determinação de geometrias moleculares onde não há um rigor na determinação de mínimos locais em ordem de energia. O NGA é eficiente na determinaçãoao do GM, realizando em um menor tempo, se comparado ao RS e IGA. O IGA mostrou-se um método mais robusto que os outros, pois além de determinar o GM é possível encontrar os mínimos locais em ordem de energia. Realizando cálculos em um tempo intermediário ao RS e NGA, o IGA determinou o GM assim como o NGA, e encontrou estruturas que não foram possíveis utilizando o RS. Os GM’s da asparagina determinados utilizando os potenciais PC1, PM3, AM1 e RM1 possuem uma grande diferença estrutural. Isto demonstra que diferentes potencias utilizados nos cálculos de estrutura eletrônica podem levar a diferentes resultados. Ao analisarmos as estruturas obtidas para os potenciais PC1, PM3, AM1 e RM1, utilizando o IGA, constata-se que há uma diferença na topologia de suas superfícies de energia potencial.

Depuis les travaux précurseurs de Richard et al., les constructions géométriques<br />occupent une place importante dans la description des macro-molécules et leurs assemblages.<br />En particulier, certains complexes cellulaires liés au diagramme de Voronoï<br />ont été utilisés pour décrire les propriétés de compacité des empilement atomiques,<br />calculer des surfaces moléculaires, ou encore détecter des cavités à la surface des molécules.<br />Cette thèse se positionne dans ce contexte, et après une brève introduction à la<br />structure des protéines, détaille quatre contributions.<br /><br />Premièrement, en utilisant le principe de balayage introduit par <br />Bentley et Ottmann, cette thèse présente le premier algorithme effectif<br />pour construire l'arrangement exact de cercles sur une sphère.<br />De plus, en supposant que les cercles proviennent de l'intersection entre<br />sphères, une stratégie pour calculer les listes couvrantes d'une face de<br />l'arrangement (i.e. la liste des boules qui la recouvrent) est proposée.<br />L'exactitude n'étant pas une fin en soi, mais plutôt une façon de rendre l'algorithmique<br />robuste, nous montrons expérimentalement que le surcoût induit est modeste.<br /><br />Deuxièmement, cette thèse développe les primitives algébriques et géométriques<br />requises par l'algorithme de balayage afin de le rendre générique et robuste. Ces primitives<br />sont intégrées dans une contexte plus général, à savoir le noyau CGAL pour les objets sphériques.<br /><br />Troisièmement, la machinerie introduite est utilisée pour traiter un problème<br />de biologie structurale computationelle : la sélection d'un sous-ensemble varié<br />à partir d'un ensemble redondant de conformations de boucles.<br />Nous proposons de résoudre ce problème de sélection en retenant les <br />représentants qui maximisent l'aire ou le volume de la sélection.<br />Ces questions peuvent être traitées géométriquement à l'aide d'arrangements de cercles sur une sphère.<br />La validation est faîte sur deux fronts.<br />D'un point de vue géométrique, nous montrons que notre approche génère des sélections<br />dont l'aire de la surface moléculaire équivaut à celle de sélections obtenues par des stratégies <br />classiques, mais qui sont de taille nettement inférieure.<br />Du point de vue amarrage de protéines, nous montrons que nos sélections améliorent de<br />manière significative les résultats obtenus à l'aide d'un algorithme manipulant des<br />parties flexibles.<br /><br />Pour finir, nous discutons les problèmes et choix d'implémentation, en<br />les replaçant dans le contexte de la librairie CGAL.

In recent years, several interesting biological roles played by RNA have come to light. Apart from their known role in translation of genetic information from DNA to protein, they have been shown to act as enzymes as well as regulators of gene expression. Protein-RNA complexes are involved in regulating cellular processes like cell division, differentiation, growth, cell aging and death. A number of clinically important viruses have RNA as their genetic material. Defective RNA molecules have been linked to a number of human diseases. The ability of RNA to adopt stunningly complex three-dimensional structures aids in diverse functions like catalysis, metabolite sensing and transcriptional control. Several secondary structure motifs are observed in RNA, of which the double-helical RNA motif is ubiquitous and well characterized. Though DNA duplexes have been shown to be present in many polymorphic states, RNA duplexes are believed to exhibit conservatism. Early fibre diffraction analysis on molecular structures of natural and synthetically available oligo- and polynucleotides suggested that the double-helical structures of RNA might exist in two forms: A-form and A′-form. New improved crystallographic methods have contributed to the increased availability of atomic resolution structures of many biologically significant RNA molecules. With the available structural information, it is feasible to try and understand the contribution of the variations at the base pair, base-pair step and backbone torsion angle level to the overall structure of the RNA duplex. Further, the effect of protein binding on RNA structure has not been extensively analysed. These studies have not been investigated in greater detail due to the focus of the research community on understanding conformational changes in proteins when bound to RNA, and due to the lack of a significant number of solved RNA structures in both free and protein-bound state. While studies on the conformation of the DNA double-helical stem have moved beyond the dinucleotide step into tri-, tetra-, hexa- and octanucleotide levels, similar knowledge for RNA even at the dinucleotide step level is lacking. In this thesis, the results of detailed analyses to understand the contribution of the base sequence towards RNA conformational variability as well as the structural changes incurred upon protein binding are reported. Objectives The primary objective of this thesis is to understand the following through detailed analyses of all available high-resolution crystal structures of RNA. 1 Exploring sequence-dependent variations exhibited by dinucleotide steps formed by Watson-Crick (WC) base pairs in RNA duplexes. 2 Identifying sequence-dependent variations exhibited by dinucleotide steps containing non-Watson-Crick (NWC) base pairs in RNA duplexes. 3 Developing a web application for the generation of sequence-dependent non-uniform nucleic acid structures. 4 Investigating the relationship between base sequence and backbone torsion-angle preferences in RNA double helices followed by molecular dynamics simulation using various force fields, to check their ability to reproduce the above experimental findings. Chapter 1 gives an overview of the structural features and polymorphic states of RNA duplexes and the present understanding of the structural architecture of RNA, thereby laying the background to the studies carried out subsequently. The chapter also gives a brief description on the methodologies applied. Relevant methodologies and protocols are dealt with in detail in the respective chapters. Sequence-dependent base-pair step geometries in RNA duplexes A complete understanding of the conformational variability seen in duplex RNA molecules at the dinucleotide step level can aid in the understanding of their function. This work was carried out to derive geometric information using a non-redundant RNA crystal structure dataset and to understand the conformational features (base pair and base-pair step parameters) involving all Watson-Crick (WC) (Chapter 2) and non-Watson-Crick (NWC) base pairs (Chapter 3). The sequence-dependent variations exhibited by the base-pair steps in RNA duplexes are elaborated. Further, potential non-canonical hydrogen bond interactions in the steps are identified and their relationship with dinucleotide step geometry is discussed. Comparison of the features of dinucleotide steps between free and protein-bound RNA datasets suggest variations at the base-pair step level on protein binding, which are more pronounced in non-Watson-Crick base pair containing steps. Chapter 4 describes a web-server NUCGEN-Plus, developed for building and regeneration of curved and non-uniform DNA and RNA duplexes. The main algorithm is a modification of our earlier program NUCGEN that worked mainly for DNA. The WC step parameters and intra-base parameters for RNA were obtained from the work detailed in Chapter 2. The FORTRAN code and input sequence file format was modified. The program has two modules: a) Using the model-building module, the program can build duplex structures for a given input DNA/RNA sequence. Options are available for selecting various derived or user specified base-pair step parameters, and fibre diffraction parameters that can be used in the building process. The program can generate double-helical structures up to 2000 nucleotides in length. In addition, the program can calculate the curvature of the generated duplex at defined length scale. b) Using the regeneration module, double-helical structures of nucleic acids can be rebuilt from the existing solved structures. Further, variants of an existing structure can be generated by varying the input geometric parameters. The web-server has a user-friendly interface and is freely available in the public domain at: http://nucleix.mbu.iisc.ernet.in/nucgenplus/index.html Sequence dependence of backbone torsion angle conformers in RNA duplexes RNA molecules consist of covalently linked nucleotide units. Each of these units has six rigid torsional degrees of freedom (α, β, γ, δ, ε, and ζ) for the backbone and one (χ) around the glycosidic bond connecting the base to the ribose, thereby providing conformational flexibility. An understanding of the relationship between base sequence and structural variations along the backbone can help deduce the rationale for sequence conservation and also their functional importance. Chapter 5 describes in detail the torsion angle-dependent variations seen in dinucleotide steps of RNA duplex. A non-redundant, high resolution (≤2.5Å) crystal structure dataset was created. Base-specific preferences for the backbone and glycosidic torsion angles were observed. Non-A-form torsion angle conformers were found to have a greater prevalence in protein-bound duplexes. Further validation of the above observation was performed by analysing the RNA backbone conformers and the effect of protein binding, in the crystal structure of E. coli 70S ribosome. Chapter 5 further describes the molecular dynamics simulation studies carried out to understand the effect of force fields on the RNA backbone conformer preferences. A 33mer long duplex was simulated using seven different force fields available in AMBER and CHARMM program, each for 100 ns. Trajectory analyses suggest the presence of sequence-dependent torsion angle preferences. Torsion angle conformer distribution closer to that of crystal structures was observed in the system simulated using parmbsc0 force field. Molecular dynamics simulation studies of AU/AU base-pair step A unique geometric feature, unlike that in other purine-pyrimidine (RY) steps in the crystal dataset analysis, was reported for AU/AU step (see Chapter 2). Appendix 1 describes the work carried out to validate these features observed in the crystal structures using simulation studies. Additionally, the effect of nearest-neighbor base pairs on the AU/AU step geometry were examined. General Conclusion Overall, the findings of this thesis work suggest that RNA duplexes exhibit sequence-dependent structural variations and sample a large volume of the double-helical conformational space. Further, protein binding affects the local base-pair step geometry and backbone conformation.

<p>Research on graph representation learning (GRL) has made major strides over the past decade, with widespread applications in domains such as e-commerce, personalization, fraud & abuse, life sciences, and social network analysis. Despite its widespread success, fundamental questions on practices employed in modern day GRL have remained unanswered. Unraveling and advancing two such fundamental questions on the practices in modern day GRL forms the overarching theme of my thesis.</p> <p>The first part of my thesis deals with the mathematical foundations of GRL. GRL is used to solve tasks such as node classification, link prediction, clustering, graph classification, and so on, albeit with seemingly different frameworks (e.g. Graph neural networks for node/graph classification, (implicit) matrix factorization for link prediction/ clustering, etc.). The existence of very distinct frameworks for different graph tasks has puzzled researchers and practitioners alike. In my thesis, using group theory, I provide a theoretical blueprint that connects these seemingly different frameworks, bridging methods like matrix factorization and graph neural networks. With this renewed understanding, I then provide guidelines to better realize the full capabilities of these methods in a multitude of tasks.</p> <p>The second part of my thesis deals with cases where modeling real-world objects as a graph is an oversimplified description of the underlying data. Specifically, I look at two such objects (i) modeling hypergraphs (where edges encompass two or more vertices) and (ii) using GRL for predicting protein properties. Towards (i) hypergraphs, I develop a hypergraph neural network which takes advantage of the inherent sparsity of real world hypergraphs, without unduly sacrificing on its ability to distinguish non isomorphic hypergraphs. The designed hypergraph neural network is then leveraged to learn expressive representations of hyperedges for two tasks, namely hyperedge classification and hyperedge expansion. Experiments show that using our network results in improved performance over the current approach of converting the hypergraph into a dyadic graph and using (dyadic) GRL frameworks. Towards (ii) proteins, I introduce the concept of conditional invariances and leverage it to model the inherent flexibility present in proteins. Using conditional invariances, I provide a new framework for GRL which can capture protein-dependent conformations and ensures that all viable conformers of a protein obtain the same representation. Experiments show that endowing existing GRL models with my framework shows noticeable improvements on multiple different protein datasets and tasks.</p>

Ce travail présente une étude théorique des spectres RMN de molécules biologiques. Dans la première partie, les calculs DFT des paramètres RMN (déplacements chimiques et constantes de couplage spin-spin) pour les protons liés à des atomes de carbone ont été réalisés pour quatre métabolites de la prostate: la putrescine, la spermidine, la spermine, et la sarcosine, et trois métabolites du cerveau: l'acétate, l'alanine et la sérine. Une étude théorique systématique, dans l'approche DFT, des paramètres de RMN des métabolites a montré que la méthode B3LYP/6-311++G** est un bon compromis entre la précision et les coûts. Les contributions du solvant ont été évaluées en utilisant le modèle PCM, les effets des isomères, pondérés dans l’approximation de Boltzmann, ont été pris en compte, et les corrections de vibration de point zéro ont été estimées en utilisant une approche perturbative au second ordre. La comparaison avec l'expérience a démontré que tous ces effets sont nécessaires pour améliorer l'accord entre les données calculées et expérimentales, aboutissant à des résultats de grande précision. Dans la deuxième partie, nous avons développé un nouveau modèle, BioShift, qui permet la prédiction des déplacements chimiques des différents noyaux (H, N, C ...) pour des molécules biologiques (protéines, ADN, ARN, polyamine ...). Il est simple, rapide, et comporte un nombre limité de paramètres. La comparaison avec des modèles sophistiqués conçus spécialement pour la prédiction des déplacements chimiques des protéines a montré que Bioshift est concurrentiel avec de tels modèles. / The present work presents a theoretical study of the NMR spectra of biological molecules. In the first part, DFT calculations of the spin-Hamiltonian NMR parameters (chemical shifts and spin-spin coupling constants) for protons attached to carbon atoms have been performed for four prostate metabolites: putrescine, spermidine, spermine, and sarcosine, and three brain metabolites: acetate, alanine, and serine. A theoretical investigation, within the DFT approach, of the NMR parameters of metabolites has shown that the B3LYP/6-311++G** level of calculation is a good compromise between accuracy and costs. Contributions from solvent were evaluated using the PCM model, Boltzmann weighted isomer effects were calculated, and zero-point vibrational corrections were estimated using a second order perturbation approach. Comparison with experiment has demonstrated that all these effects are necessary to improve the agreement between calculated and experimental data. In the second part, we have presented a new model, BioShift, that allows the prediction of chemical shifts of different nuclei (H, N, C…) for biological molecules (proteins, DNA, RNA, polyamine …). It is simple, fast, and involves a limited number of parameters. Comparison with well-known sophisticated models designed especially for the prediction of chemical shifts of proteins showed that Bioshift is competitive with such models.