Proteochemometrics (PCM) est une bioactivité prophétique la méthode posante de simultanément modeler la bioactivité de ligands multiple contre des objectifs multiples... / Proteochemometrics (PCM) is a predictive bioactivity modelling method to simultaneously model the bioactivity of multiple ligands against multiple targets. Therefore, PCM permits to explore the selectivity and promiscuity of ligands on biomolecular systems of different complexity, such proteins or even cell-line models. In practice, each ligand-target interaction is encoded by the concatenation of ligand and target descriptors. These descriptors are then used to train a single machine learning model. This simultaneous inclusion of both chemical and target information enables the extra- and interpolation to predict the bioactivity of compounds on targets, which can be not present in the training set. In this thesis, a methodological advance in the field is firstly introduced, namely how Bayesian inference (Gaussian Processes) can be successfully applied in the context of PCM for (i) the prediction of compounds bioactivity along with the error estimation of the prediction; (ii) the determination of the applicability domain of a PCM model; and (iii) the inclusion of experimental uncertainty of the bioactivity measurements. Additionally, the influence of noise in bioactivity models is benchmarked across a panel of 12 machine learning algorithms, showing that the noise in the input data has a marked and different influence on the predictive power of the considered algorithms. Subsequently, two R packages are presented. The first one, Chemically Aware Model Builder (camb), constitues an open source platform for the generation of predictive bioactivity models. The functionalities of camb include : (i) normalized chemical structure representation, (ii) calculation of 905 one- and two-dimensional physicochemical descriptors, and of 14 fingerprints for small molecules, (iii) 8 types of amino acid descriptors, (iv) 13 whole protein sequence descriptors, and (iv) training, validation and visualization of predictive models. The second package, conformal, permits the calculation of confidence intervals for individual predictions in the case of regression, and P values for classification settings. The usefulness of PCM to concomitantly optimize compounds selectivity and potency is subsequently illustrated in the context of two application scenarios, which are: (a) modelling isoform-selective cyclooxygenase inhibition; and (b) large-scale cancer cell-line drug sensitivity prediction, where the predictive signal of several cell-line profiling data is benchmarked (among others): basal gene expression, gene copy-number variation, exome sequencing, and protein abundance data. Overall, the application of PCM in these two case scenarios let us conclude that PCM is a suitable technique to model the activity of ligands exhibiting uncorrelated bioactivity profiles across a panel of targets, which can range from protein binding sites (a), to cancer cell-lines (b).

Le carcinome hépatocellulaire (HCC) est le cancer du foie le plus répandu et représente la seconde cause de décès par cancer dans le monde. Un mauvais pronostic et l'absence de traitement efficace en font un problème majeur de santé publique dans les pays en voie de développement, notamment en Asie du Sud-Est, justifiant pleinement la recherche de molécules ou d'approches thérapeutiques nouvelles contre l'HCC. Ce travail porte sur la recherche de molécules isolées de plantes vietnamiennes actives contre l'HCC. La première approche a consisté en un criblage pharmacologique de 33 substances naturelles qui a conduit à l'identification de 7 ent-kaurane diterpénoïdes isolés de Croton kongensis Gagnep. présentant des propriétés antiprolifératives originales. La seconde approche, par criblage in silico d'une banque de 354 substances naturelles, a permis d'identifier la solasonine comme inhibiteur de l'interaction mortalin - p53 induisant l'apoptose dans la lignée cellulaire humaine HepG2. / Human hepatocellular carcinoma (HCC) is the most common type of liver cancer, the second most common cause of death from cancer worldwide. A very poor prognosis and a lack of effective treatments make liver cancer a major public health problem, notably in less developed regions, particularly in Eastern Asia. This fully justifies the search of new molecules and therapeutic strategies against HCC. The present work focused on finding bioactive compounds from Vietnamese plants against HCC. The first approach used classical screening of 33 natural compounds which resulted in the identification of 7 ent-kaurane diterpenoids isolated from Croton kongensis Gagnep. as potential agents. The second approach aimed at identifying molecules that could abrogate the interaction between Mortalin and p53 by in silico screening of a database of 354 natural compounds, which allowed the identification of Solasonine as a potent inhibitor of p53 - mortalin interactions.

As doenças negligenciadas são conseqüências marcantes do subdesenvolvimento que atinge diversas regiões do planeta. Dentre estas, destaca-se a Doença de Chagas, causada pelo parasita Trypanosoma cruzi, a qual afeta aproximadamente um quarto da população da América Latina e para qual os fármacos utilizados apresentam baixa eficácia, toxidez e sérios efeitos colaterais. Este quadro é agravado pela emergência de cepas resistentes, o que indica a grande necessidade de desenvolvimento de novos agentes quimioterápicos contra esta doença. A enzima gliceraldeído-3-fosfato desidrogenase (GAPDH) da via glicolítica do T. cruzi é um alvo macromolecular interessante devido ao seu papel essencial no metabolismo de tripanossomatídeos. Constitui o objetivo desta tese o desenvolvimento, a padronização e a validação de ensaios enzimáticos para a realização de extensivas triagens bioquímicas de modo a contribuir para a identificação de novos inibidores da GAPDH pertencentes a diversas classes químicas. Os compostos estudados são provenientes de síntese orgânica e de complexos inorgânicos de Rutênio, bem como de origem natural. Paralelamente, a realização do ensaio virtual em larga escala a partir de uma base de dados dirigida e da estrutura da GAPDH de T. cruzi resultou na identificação de um inibidor inédito da proteína alvo. Estudos do mecanismo de ação enzimático levaram à elucidação da modalidade de alguns inibidores identificados nesse estudo. / Parasitic diseases are a major global cause of illness, long-term disability, and death, with severe socio-economic consequences for millions of people worldwide. In Latin America, nearly one fourth of the population is infected by Trypanosoma cruzi, the causative agent of Chagas disease. The limited existing drug therapies suffer from a combination of drawbacks including poor efficacy, resistance and serious side effects. Therefore, there is an urgent need for new drugs that can overcome resistance and are safe and effective for use in human. The crucial dependence on glycolysis as a source of energy makes the glycolytic parasite enzymes promising targets for drug design. In this context, the enzyme glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was identified as an attractive target for drug design. The development of standard enzymatic assays combined with extensive biological screening was the aim of this work and it has been contributing for the identification of novel GAPDH inhibitors. These compounds belong to several chemical classes from different sources, including organic synthesis, inorganic complexes and natural products. In addition, a virtual screening approach was applied in a focused database, previously filtered by drug-like properties, in order to identify new hits. This strategy resulted in the discovery of a novel scaffold with significant inhibitory activity against T. cruzi GAPDH. Kinetic and inhibition assays were conducted to shed light on the mechanism of action of the promising inhibitors.

Appartenant à la sous-famille des récepteurs couplés aux protéines G (RCPGs) apparentés àla rhodopsine et identifiés comme des cibles à fort potentiel thérapeutique, les récepteurs MT, et MT2à la mélatonine, 5-HT2C à la sérotonine et CB2 aux cannabinoïdes ont été étudiés par des approches insilico afin de mettre en évidence les déterminants structuraux critiques pour l'affinité, la sélectivité etl'activité pharmacologique de leurs ligands. Bénéficiant de données cristallographiques récentes,plusieurs états conformationnels de ces quatre récepteurs ont été modélisés en fonction du profilpharmacologique recherché. L'étude comparative de ces différents états conformationnels par dessimulations de dynamique moléculaire a permis de caractériser le rôle prépondérant joué par la boucleextracellulaire E2 et l'hélice 6 dans les mécanismes d'activation de ces RCPGs. Sur la base deméthodes chémoinformatiques, le criblage virtuel de ligands ciblant ces modèles tridimensionnels apermis de caractériser un modèle du récepteur 5-HT2C très spécifique de ligands agonistes inverses etd'identifier des touches pharmacologiques sur les récepteurs MTi et CB2. / Identified as highly relevant therapeutical targets, the MT, and MT2 melatonin receptors, the5-HT2C serotonin and the CB2 cannabinoid receptors, which belong to the rhodopsin-like G proteincoupledreceptors (GPCRs) subfamily, have been studied by in silico approaches in order to identifycritical structural features for the binding, the selectivity and the pharmacological activity of theirligands. Gaining by sottie recent crystallographic data, various conformational states of these fourreceptors have been modeled according to the expected pharmacological profile. The comparativestudy of these various conformational states by molecular dynamics simulations has led to emphasizethe crucial rôle of the E2 extracellular loop and hélix 6 in the activation mechanisms of these GPCRs.On the basis of chemoinformatic methods, the virtual ligand screening targeting these threedimensionalmodels has promoted the characterization of a 5-HT2C receptor model able to bindspecifically inverse agonist ligands and the identification of pharmacological hits targeting the MTiand CB2 receptors.

Drug discovery is a process that takes many years and hundreds of millions of dollars to reveal a confident conclusion about a specific treatment. Part of this sophisticated process is based on preliminary investigations to suggest a set of chemical compounds as candidate drugs for the treatment. Computational resources have been playing a significant role in this part through a step known as virtual screening. From a data mining perspective, availability of rich data resources is key in training prediction models. Yet, the difficulties imposed by big expansion in data and its dimensionality are inevitable. In this thesis, I address the main challenges that come when data mining techniques are used for virtual screening. In order to achieve an efficient virtual screening using data mining, I start by addressing the problem of feature selection and provide analysis of best ways to describe a chemical compound for an enhanced screening performance. High-throughput screening (HTS) assays data used for virtual screening are characterized by a great class imbalance. To handle this problem of class imbalance, I suggest using a novel algorithm called DRAMOTE to narrow down promising candidate chemicals aimed at interaction with specific molecular targets before they are experimentally evaluated. Existing works are mostly proposed for small-scale virtual screening based on making use of few thousands of interactions. Thus, I propose enabling large-scale (or big) virtual screening through learning millions of interaction while exploiting any relevant dependency for a better accuracy. A novel solution called DRABAL that incorporates structure learning of a Bayesian Network as a step to model dependency between the HTS assays, is showed to achieve significant improvements over existing state-of-the-art approaches.

Virtual screening is a well-established part of computer-aided drug design, which heavily employs similarity search and similarity modeling methods. Most of the popular methods are target agnostic, leaving space for design of new methods that would take into account the specifics of the particular molecular target. Additionally, newly developed methods suffer from two related issues: benchmarking and availability. Benchmarking in the domain often suffers from the use of inappropriate reference methods, lack of reproducibility, and the use of nonstandard benchmark datasets. Although there have been several benchmarking studies in the domain that aim at addressing these issues, mainly by offering a standardized comparison, they often suffer from similar drawbacks. For these reasons, new methods fail to gain trust and therefore fail to become a part of the standard toolbox, which thus consists mostly of older methods. In this work, we address the above-described issues. First, we introduce new adaptive methods for virtual screening. Then, to make our and other newly developed methods readily available, we have designed and implemented a virtual screening tool. To address the benchmarking issue, we have compiled a publicly available collection of benchmarking datasets and proposed a platform offering a...

Magister Pharmaceuticae - MPharm / Malaria disease poses substantial health risks to many nations, especially in Africa, where it primarily affects pregnant women, children, and immunocompromised patients. However, current antimalarial drugs have limitations such as low safety profile and particularly widespread treatment failure due to the increasing resistance of Plasmodium falciparum, the major causative organism to artemisinin-based therapy (ACT) and other chemotherapeutics. In the light of this, there is a pressing need for new antimalarial drugs with novel mechanisms of action and satisfactory pharmacokinetic properties, which has led to the current study. Furthermore, current antimalarial drugs target specific stages of the Plasmodium life cycle. For instance, chloroquine targets the erythrocytic stage while primaquine targets the liver stage. However, these therapies cannot achieve complete elimination of the parasite once the life cycle has been established in the body. Hence, the goal of this study is to combat resistance by finding novel compounds that can bind to a multiple-staged protein in Plasmodium falciparum. Based on this consideration, falstatin was chosen as the protein target for this study because it was observed to play a crucial role in the degradation of haemoglobin, rupture of erythrocytes by mature schizonts, and subsequent invasion of erythrocytes by free merozoites. Hence, the protein, falstatin can be targeted to inhibit cell growth and cause plasmodial cell death in merozoites as well as schizonts of Plasmodium falciparum. Therefore, it is intended that compounds that bind to falstatin could serve as novel antimalarials that target multiple stages of the Plasmodium life cycle. Consequently, this study explored the structure-based virtual screening approach to identify compounds that could bind to the protein target, falstatin in Plasmodium falciparum. An extensive literature review identified falstatin as the multi-staged drug target for this study, while homology modelling was used to generate the three-dimensional structure of falstatin. Molecular docking was conducted to predict the binding energy of compiled antiplasmodial compounds to falstatin while druglikeness analysis was used to prioritize compounds according to their ADMET (absorption, distribution, metabolism, excretion and toxicity) properties. The top-ranked compound, based on a novel ligand scoring function, was then subjected to molecular dynamics (MD). Following this step, rescoring analysis was performed on the top 5 compounds using the Molecular Mechanics-Generalized Born Surface Area (MM-GBSA) scoring function to gain insight into their component binding energies. Thereafter, a pharmacophore hypothesis was developed based on the 5 top-ranking compounds in order to screen other compound libraries in the future. From the results, TCMDC 131646, TCMDC-124274, TCMDC-138266, TCMDC 123844 and TCMDC 131234 possessed good binding energies and satisfactory ADMET properties showing high ligand scores of 77.1, 75.4, 75.4, 75.4 and 73.1 respectively (on a total scale of 100). Also, the study revealed that the top-ranked compound, TCMDC 131646 had a binding energy of -6.15 KJ/mol, contained no toxicophore and conformed to Lipinski, Egan and Muegge rules of druglikeness. Findings from the MD simulation demonstrated that TCMDC 131646 strongly interacted with the protein, falstatin. Morealso, the study revealed that TCMDC 131646 is structurally diverse from chloroquine, artemisinin, artemether and lumefantrine, indicating that it may possess a distinct mechanism of action. The rescoring analysis of TCMDC-131646, TCMDC 124274, TCMDC-138266, TCMDC 123844 and TCMDC 131234 predicted negative binding energies ≤ -4.662 KJ/mol for the top compounds, further indicating that these compounds are likely to bind strongly with falstatin. Additionally, the developed pharmacophore hypothesis contained -H-N-C=O and N-H moieties which strongly suggested that the presence of electron-withdrawing groups could be vital for the inhibition of falstatin at the active site. Overall, TCMDC 131646 was predicted to be a drug-like and safe compound that could inhibit falstatin in Plasmodium falciparum. Chemical-disease co-occurrence analysis in literature revealed that this compound showed in-vitro antiplasmodial activity at an IC50 of 0.226μM and has also shown in vitro activity for neuralgia, hyperalgesia and arthritis. The research recommends TCMDC 131646 as a potential antimalarial hit compound that could yield novel analogues by hit expansion. However, confirmatory in-vitro and in-vivo studies are required to substantiate these predictions

No description available.

Lanthionine synthetase C-like protein 2 (LANCL2) is a member of the LANCL protein family, which is broadly expressed throughout the body. LANCL2 is the molecular target of abscisic acid (ABA), a compound with insulin-sensitizing and immune modulatory actions. LANCL2 is required for membrane binding and signaling of ABA in immune cells. Direct binding of ABA to LANCL2 was predicted in silico using molecular modeling approaches and validated experimentally using ligand-binding assays and kinetic surface plasmon resonance studies. The therapeutic potential of the LANCL2 pathway ranges from increasing cellular sensitivity to anticancer drugs, insulin-sensitizing effects and modulating immune and inflammatory responses in the context of immune-mediated and infectious diseases. A case for LANCL2-based drug discovery and development is also illustrated by the anti-inflammatory activity of novel LANCL2 ligands such as NSC61610 against inflammatory bowel disease in mice. This dissertation discusses the value of LANCL2 as a novel therapeutic target for the discovery and development of new classes of orally active drugs against chronic metabolic, immune-mediated and infectious diseases and as a validated target that can be used in precision medicine. Specifically, in Chapter 2 of the dissertation, we performed homology modeling to construct a three-dimensional structure of LANCL2 using the crystal structure of LANCL1 as a template. Our molecular docking studies predicted that ABA and other PPAR - agonists share a binding site on the surface of LANCL2. In Chapter 3 of the dissertation, structure-based virtual screening was performed. Several potential ligands were identified using molecular docking. In order to validate the anti-inflammatory efficacy of the top ranked compound (NSC61610) in the NCI Diversity Set II, a series of in vitro and pre-clinical efficacy studies were performed using a mouse model of dextran sodium sulfate (DSS)-induced colitis. In Chapter 4 of the dissertation, we developed a novel integrated approach for creating a synthetic patient population and testing the efficacy of the novel pre-clinical stage LANCL2 therapeutic for Crohn's disease in large clinical cohorts in silico. Efficacy of treatments on Crohn's disease was evaluated by analyzing predicted changes of Crohn's disease activity index (CDAI) scores and correlations with immunological variables were evaluated. The results from our placebo-controlled, randomized, Phase III in silico clinical trial at 6 weeks following the treatment shows a positive correlation between the initial disease activity score and the drop in CDAI score. This observation highlights the need for precision medicine strategies for IBD. / Ph. D.

Peroxisome proliferator-activated receptor γ (PPARγ) is associated with a wide range of diseases, including type 2 diabetes mellitus (T2D). Thiazolidinediones (TZDs) are agonists of PPARγ which have an insulin sensitizing effect, and are therefore used as a treatment for T2D. However, TZDs cause negative side effects in patients, such as weight gain, edema, and increased risk of bone fracture. Partial agonists could be an alternative to TZD-based drugs with fewer side effects. However, there is a lack of understanding of the types of PPARγ partial agonists and how they differ from full agonists. In silico techniques, like virtual screening, molecular docking, and pharmacophore modeling, allow us to determine and characterize markers of varying levels of agonism. An extensive search of the RCSB Protein Data Bank found 62 structures of PPARγ resolved with partial agonists. Cross-docking was performed and found that two PDB structures, 3TY0 and 5TWO, would be effective as receptor structures for virtual screening. By clustering known partial agonists by common pharmacophore features, we found several distinct groups of partial agonists. Interaction and pharmacophore models were created for each group of partial agonists. Virtual screening of FDA-approved compounds showed that the models were able to predict potential partial agonists of PPARγ. This study provides additional insight into the different binding modes of partial agonists of PPARγ and their characteristics. These models can be used to assist drug discovery efforts for intelligently designing novel therapeutics for T2D which have fewer negative side effects. / Master of Science in Life Sciences / The peroxisome proliferator-activated receptor γ (PPARγ) protein is associated with a wide range of diseases, including type 2 diabetes mellitus (T2D). Thiazolidinediones (TZDs) are compounds that activate PPARγ, and increase insulin sensitivity in patients with T2D. However, TZDs cause negative side effects in patients, such as weight gain, increased fluid retention, and increased risk of bone fracture. Partial agonists could be an alternative to TZD-based drugs with fewer side effects. However, there is a lack of understanding of the types of PPARγ partial agonists and how they differ from full agonists. Computational techniques allow us to investigate common features between known partial agonists. An extensive search of the RCSB Protein Data Bank found 62 structures of PPARγ which contained partial agonists. Each known partial agonist was docked into twelve complete PPARγ structures, and it was found that two structure models would be effective as receptor structures for virtual screening. A set of known partial agonists were grouped based on common chemical features, and three distinct groups of partial agonists were found. Binding criteria for each of these three groups were developed. A library of FDA-approved compounds was screened using the criteria for binding to identify potential novel partial agonists. Three potential novel partial agonists were found in the screening. This study provides additional insight into how different compounds activate PPARγ. These methods can be used to assist drug discovery efforts for intelligently designing novel therapeutics for T2D which have fewer negative side effects.