This work is based on the Molpher SW project, which is client-server application aiding exploration of chemical space between two input molecules. Aim of master thesis was modify the current version of program to manage scaffold hopping technique. This technique represents molecule in a simplified way. The simpler molecule is called scaffold. First of all there was need to define seve- ral levels of granularity and for each level define morphing operators. Server was modified with respect for parallelization. Experimental exploration of chemical space with and without the new feature is part of this work too. Powered by TCPDF (www.tcpdf.org)

The purpose of this thesis was to design and implement a hierarchical approach to visualization of the chemical space. Such visualization is a challenging yet important topic used in diverse fields ranging from material engineering to drug design. Especially in drug design, modern methods of high- throughput screening generate large amounts of data that would benefit from hierarchical analysis. One possible approach to hierarchical classification of molecules is a structure based classification based on molecular scaffolds. The scaffolds are widely used by medicinal chemists to group molecules of similar properties. A few scaffold-based hierarchical visualization methods have been proposed. However, to our best knowledge, there exists no tool that would provide a scaffold-based hierarchical visualization of molecular data sets on the background of known chemical space. In this thesis, such tool was created. First, a scaffold tree hierarchy based on ring topologies was designed. Next, this hierarchy was used to analyze frequency of scaffolds extracted from molecules in PubChem Compound database. Subsequently, the PubChem Compound scaffold frequency data was used as a background for visualization of molecular data sets. The visualization is performed by a client-server application implemented as a part of...

De nouveaux domaines ont vu le jour, à l’interface entre biologie, chimie et informatique, afin de répondre aux multiples problématiques liées à la recherche de médicaments. Cette thèse se situe à l’interface de plusieurs de ces domaines, regroupés sous la bannière de la chémo-informatique. Récent à l’échelle humaine, ce domaine fait néanmoins déjà partie intégrante de la recherche pharmaceutique. De manière analogue à la bioinformatique, son pilier fondateur reste le stockage, la représentation, la gestion et l’exploitation par ordinateur de données provenant de la chimie. La chémoinformatique est aujourd’hui utilisée principalement dans les phases amont de la recherche de médicaments. En combinant des méthodes issues de différents domaines (chimie, informatique, mathématique, apprentissage, statistiques, etc.), elle permet la mise en oeuvre d’outils informatiques adaptés aux problématiques et données spécifiques de la chimie, tels que le stockage de l’information chimique en base de données, la recherche par sous-structure, la visualisation de données, ou encore la prédiction de propriétés physico-chimiques et biologiques.Dans ce cadre pluri-disciplinaire, le travail présenté dans cette thèse porte sur deux aspects importants liés à la chémoinformatique : (1) le développement de nouvelles méthodes permettant de faciliter la visualisation, l’analyse et l’interprétation des données liées aux ensembles de molécules, plus communément appelés chimiothèques, et (2) le développement d’outils informatiques permettant de mettre en oeuvre ces méthodes. / Some news areas in biology ,chemistry and computing interface, have emerged in order to respond the numerous problematics linked to the drug research. This is what this thesis is all about, as an interface gathered under the banner of chimocomputing. Though, new on a human scale, these domains are nevertheless, already an integral part of the drugs and medicines research. As the Biocomputing, his fundamental pillar remains storage, representation, management and the exploitation through computing of chemistry data. Chimocomputing is now mostly used in the upstream phases of drug research. Combining methods from various fields ( chime, computing, maths, apprenticeship, statistics, etc…) allows the implantation of computing tools adapted to the specific problematics and data of chime such as chemical database storage, understructure research, data visualisation or physoco-chimecals and biologics properties prediction.In that multidisciplinary frame, the work done in this thesis pointed out two important aspects, both related to chimocomputing : (1) The new methods development allowing to ease the visualization, analysis and interpretation of data related to set of the molecules, currently known as chimocomputing and (2) the computing tools development enabling the implantation of these methods.

De nouveaux antiviraux à large spectre, agissant comme intercalant d'acides nucléiques, ont été identifiés par criblage virtuel et grâce à des cartes de l’espace chimique. La 1ère partie de la thèse présente le modèle QSPR pour la solubilité aqueuse des molécules organiques dans une grande gamme de températures. Ce modèle a été utilisé pour l'évaluation de la solubilité des composés antiviraux. Dans la 2ème partie de cette thèse, les filtres structuraux, les modèles QSAR et pharmacophores sont présentés. Leur utilisation pour cribler une base de données contenant plus de 3,2 M de composés est ensuite décrite. Cette étape a conduit à sélectionner 55 touches qui ont été synthétisées et testées expérimentalement. Parmi eux, deux composés ont révélé une activité élevée contre le Vaccinia virus et une faible toxicité. Dans la 3ème partie de la thèse, l'approche par Cartes Topographiques Génératives (GTM) a été utilisée pour construire des cartes 2D de l'espace chimique des composés antiviraux. Les structures chimiques et les données expérimentales des composés antiviraux, présents dans la base de données ChEMBL, ont été extraites de la base, examinées et annotées avec les principaux Genus des virus. Ce jeu de données a été utilisé pour construire des cartes sur lesquels tous les autres composés de la base de données ChEMBL ont été projetés. L'analyse de ces cartes révèle des motifs structuraux caractérisant des types particuliers d'antiviraux. / Virtual screening and cartography of chemical space approaches have been used for design of broad-spectrum antivirals acting as nucleic acids intercalators. The 1st part of thesis reports QSPR model for aqueous solubility of organic molecules within the wide temperature range. This model was later used for solubility assessment of antiviral compounds. In the second part of work, structural filters, QSAR and pharmacophore models were developed then used to screen a database containing some 3.2 M compounds. This resulted in 55 hits which were synthesized and experimentally tested. Two lead compounds displayed high activity against Vaccinia virus and low toxicity. In the 3d part of the thesis, Generative Topographic Mapping (GTM) approach was used to build 2D maps of chemical space of antiviral compounds. Experimental data on antiviral compounds were extracted from ChEMBL database, curated and annotated by major virus Genus. Selected dataset was used to build maps on which all other ChEMBL compounds were projected. Analysis of the maps revealed structural motifs characterizing particular types of antivirals.

Cette thèse est consacrée à l’analyse de l’espace chimique des composés antipaludiques. L’analyse est faite à l’aide de la méthode des cartes topographiques génératrices (GTM). Un nouveau concept des cartes universelles est introduit et discuté en détail dans cette thèse : ce sont des cartes qui sont capables d’accommoder plusieurs jeux de données et les propriétés associées simultanément. Trois types des cartes sont construits et analysés : les cartes locales, globales et universelles. Elles sont toutes compétentes à la prédiction des composés actifs contre le parasite, ainsi qu’à l’analyse de l’espace chimique. Elles nous permettent d’étudier le recouvrement des données issues des sources différentes, de détecter des terra incognita de l’espace chimique, identifier des zones correspondantes aux différents mécanismes d’action, et révéler des incohérences d’annotations des données. / This thesis is dedicated to the concept of the analysis of chemical space, and the application of thatconcept to antimalarial compounds. The analysis of the chemical space of antimalarial compoundshere is done with the aid of the Generative Topographic Mapping (GTM) method. A concept ofUniversal GTM maps is developed and discussed in detail in this thesis: these are maps that areable to accommodate different datasets and associated properties. Three types of maps are builtand analyzed: local, global, and universal. All these maps perform well in predicting compoundsactive against the parasite, as well as in the analysis of chemical space: they help us to study theoverlap of data coming from different sources, detect terra incognita of the antimalarial space,delineate zones corresponding to various mechanisms of action, as well as highlight theinconsistencies in data annotations.

Cette thèse est dédiée à l’analyse systématique de l’espace chimique, et des relations structure-activité (SAR) en particulier. L’ouvrage présente des nouveaux protocoles d’analyse combinant des méthodes classiques et originales, dans le but d’analyser les SAR à l’échelle globale ainsi que locale. L’analyse globale des espaces chimiques repose sur la recherche des motifs structuraux privilégiés par cartographie topographique générative (GTM), ainsi que par analyse classique des « châssis » moléculaires. La cartographie a été ensuite couplée avec l’analyse de réseaux chimiques (CSN), permettant une transition de la vue globale vers l’analyse locale de SAR. L’optimisation mutiobjectif des propriétés de potentiels médicaments a été adressé par la méthode « star coordinates ». L’analyse locale des SAR inclut des nouvelles stratégies pour prédire les discontinuités dans le paysage structure-activité biologique, et une étude de l’impact de la structure sur l’ionisation des molécules. Des matrices SAR ont servi pour monitorer le progrès dans l’optimisation de nouveaux principes actifs. / This thesis presents studies devoted to aid in systematic analysis of chemical spaces, focusing on mining and visualization of structure-activity relationships (SARs). It reports some new analysis protocols, combining both existing and on-purpose developed novel methodology to address both large-scale and local SAR analysis. Large-scale analysis featured both generative topographic mapping (GTM)-based extraction of privileged structural motifs and scaffold analysis. GTM was combined with chemical space network (CSN) to develop a visualization tool providing global-local views of SAR in large data sets. We also introduce star coordinates (STC) to visualize multi-property space and prioritize drug-like subspaces. Local SAR monitoring includes new strategies to predict activity cliffs using support vector machine models and a study of structural modifications on ionization state of compounds. The SAR matrix methodology was applied to objectively evaluate SAR progression during lead optimization.

Starting from the observation that substances and reactions are the central entities of chemistry, I have structured chemical knowledge into a formal space called a directed hypergraph, which arises when substances are connected by their reactions. I call this hypernet chemical space. In this thesis, I explore different levels of description of this space: its evolution over time, its curvature, and categorical models of its compositionality. The vast majority of the chemical literature focuses on investigations of particular aspects of some substances or reactions, which have been systematically recorded in comprehensive databases such as Reaxys for the last 200 years. While complexity science has made important advances in physics, biology, economics, and many other fields, it has somewhat neglected chemistry. In this work, I propose to take a global view of chemistry and to combine complexity science tools, modern data analysis techniques, and geometric and compositional theories to explore chemical space. This provides a novel view of chemistry, its history, and its current status. We argue that a large directed hypergraph, that is, a model of directed relations between sets, underlies chemical space and that a systematic study of this structure is a major challenge for chemistry. Using the Reaxys database as a proxy for chemical space, we search for large-scale changes in a directed hypergraph model of chemical knowledge and present a data-driven approach to navigate through its history and evolution. These investigations focus on the mechanistic features by which this space has been expanding: the role of synthesis and extraction in the production of new substances, patterns in the selection of starting materials, and the frequency with which reactions reach new regions of chemical space. Large-scale patterns that emerged in the last two centuries of chemical history are detected, in particular, in the growth of chemical knowledge, the use of reagents, and the synthesis of products, which reveal both conservatism and sharp transitions in the exploration of the space. Furthermore, since chemical similarity of substances arises from affinity patterns in chemical reactions, we quantify the impact of changes in the diversity of the space on the formulation of the system of chemical elements. In addition, we develop formal tools to probe the local geometry of the resulting directed hypergraph and introduce the Forman-Ricci curvature for directed and undirected hypergraphs. This notion of curvature is characterized by applying it to social and chemical networks with higher order interactions, and then used for the investigation of the structure and dynamics of chemical space. The network model of chemistry is strongly motivated by the observation that the compositional nature of chemical reactions must be captured in order to build a model of chemical reasoning. A step forward towards categorical chemistry, that is, a formalization of all the flavors of compositionality in chemistry, is taken by the construction of a categorical model of directed hypergraphs. We lifted the structure from a lineale (a poset version of a symmetric monoidal closed category) to a category of Petri nets, whose wiring is a bipartite directed graph equivalent to a directed hypergraph. The resulting construction, based on the Dialectica categories introduced by Valeria De Paiva, is a symmetric monoidal closed category with finite products and coproducts, which provides a formal way of composing smaller networks into larger in such a way that the algebraic properties of the components are preserved in the resulting network. Several sets of labels, often used in empirical data modeling, can be given the structure of a lineale, including: stoichiometric coefficients in chemical reaction networks, reaction rates, inhibitor arcs, Boolean interactions, unknown or incomplete data, and probabilities. Therefore, a wide range of empirical data types for chemical substances and reactions can be included in our model.

<p>Molecular interactions lie at the heart of myriad biological processes. Knowledge of molecular recognition processes and the ability to model and predict interactions of any biological molecule to any chemical compound are the key for better understanding of cell functions and discovery of more efficacious medicines.</p><p>This thesis presents contributions to the development of a novel chemo-bioinformatics approach called proteochemometrics; a general method for interaction space analysis of biological macromolecules and their ligands. In this work we explore proteochemometrics-based interaction models over broad groups of protein families, evaluate their validity and scope, and compare proteochemometrics to traditional modeling approaches.</p><p>Through the proteochemometric analysis of large interaction data sets of multiple retroviral proteases from various viral species we investigate complex mechanisms of drug resistance in HIV-1 and discover general physicochemical determinants of substrate cleavage efficiency and binding in retroviral proteases. We further demonstrate how global proteochemometric models can be used for design of protease inhibitors with broad activity on drug-resistant viral mutants, for monitoring drug resistance mechanisms in the physicochemical sense and prediction of potential HIV-1 evolution trajectories. We provide novel insights into the complexity of HIV-1 protease specificity by constructing a generalized IF-THEN rule model based on bioinformatics analysis of the largest set of HIV-1 protease substrates and non-substrates.</p><p>We discuss how proteochemometrics can be used to map recognition sites of entire protein families in great detail and demonstrate how it can incorporate target variability into drug discovery process. Finally, we assess the utility of the proteochemometric approach in evaluation of ADMET properties of drug candidates with a special focus on inhibition of cytochrome P450 enzymes and investigate application of the approach in the pharmacogenomics field.</p>

Molecular interactions lie at the heart of myriad biological processes. Knowledge of molecular recognition processes and the ability to model and predict interactions of any biological molecule to any chemical compound are the key for better understanding of cell functions and discovery of more efficacious medicines. This thesis presents contributions to the development of a novel chemo-bioinformatics approach called proteochemometrics; a general method for interaction space analysis of biological macromolecules and their ligands. In this work we explore proteochemometrics-based interaction models over broad groups of protein families, evaluate their validity and scope, and compare proteochemometrics to traditional modeling approaches. Through the proteochemometric analysis of large interaction data sets of multiple retroviral proteases from various viral species we investigate complex mechanisms of drug resistance in HIV-1 and discover general physicochemical determinants of substrate cleavage efficiency and binding in retroviral proteases. We further demonstrate how global proteochemometric models can be used for design of protease inhibitors with broad activity on drug-resistant viral mutants, for monitoring drug resistance mechanisms in the physicochemical sense and prediction of potential HIV-1 evolution trajectories. We provide novel insights into the complexity of HIV-1 protease specificity by constructing a generalized IF-THEN rule model based on bioinformatics analysis of the largest set of HIV-1 protease substrates and non-substrates. We discuss how proteochemometrics can be used to map recognition sites of entire protein families in great detail and demonstrate how it can incorporate target variability into drug discovery process. Finally, we assess the utility of the proteochemometric approach in evaluation of ADMET properties of drug candidates with a special focus on inhibition of cytochrome P450 enzymes and investigate application of the approach in the pharmacogenomics field.

Natural products provide some of the most potent anticancer agents and offer a template for new drug design or improvement with the advantage of an enormous chemical space. The overall goal of this thesis research is to enhance the chemical space of two natural products in order to generate novel drugs with better in vivo bioactivities than the original natural products. Polycarcin V (PV) is a gilvocarcin-type antitumor agent with similar structure and comparable bioactivity with the principle compound of this group, gilvocarcin V (GV). Modest modifications of the polyketide-derived tetracyclic core of GV had been accomplished, but the most challenging part was to modify the sugar moiety. In order to solve this problem, PV was used as an alternative lead-structure for modification because its sugar moiety offered the possibility of enzymatic O-methylation. We produced four PV derivatives with different methylation patterns for cytotoxicity assays and provided important structure-activity-relationship information. Mithramycin (MTM) is the most prominent member of the aureolic acid type anticancer agents. Previous work in our laboratory generated three MTM analogues, MTM SA, MTM SK, and MTM SDK by inactivating the mtmW gene. We developed new MTM analogues by coupling many natural and unnatural amino acids to the C-3 side chain of MTM SA via chemical semi-synthesis and successfully made some compounds with both improved bioactivity and in vivo tolerance than MTM. Some of them were consequently identified as promising lead-structures against Ewing’s sarcoma. The potential of selectively generating novel MTM analogues led us to focus on a key enzyme in the biosynthetic pathway of mithramycin, MtmC. This protein is a bifunctional enzyme involved in the biosynthesis of TDP-D-olivose and TDP-D-mycarose. We clarified its enzymatic mechanisms by X-ray diffraction of several crystal complexes of MtmC with its biologically relevant ligands. Two more important post-PKS tailoring enzymes involved in the biosynthesis of the MTM side chains, MtmW and MtmGIV, are currently under investigation. This would not only give us insight into this biosynthetic pathway but also pave the way to develop potentially useful MTM analogues by engineered enzymes.